INTRODUCTION

Mylroie's (2018) objective was “…to examine the arguments concerning the Cow and Bull [megaboulders on Eleuthera] and the fenestral porosity of the Bahama Archipelago” (The Bahamas and Turks and Caicos Islands, collectively termed “BTC”). Mylroie (2018) sought to debunk several categories of evidence that support superstorms at the close of the last interglacial (e.g., Hansen et al., 2016; Hearty, 1997; Neumann and Hearty, 1996). Our “trilogy,” defined in Hearty and Tormey (2017, p. 347), identifies three lines of regional sedimentary features, (1) wave-transported megaboulders, (2) lowland chevron storm ridges, and (3) hillside runup deposits, which collectively provide evidence of frequent, intense storms in the BTC at the end of the last interglacial (marine isotope stage [MIS] 5e).

Based on limited field data from the contested sites, Mylroie's (2018) challenges included: rip-up clasts and fenestral porosity that he attributed to rainfall slurries, use of the term “washovers” (we defined as “runup,” not washovers), and evidence of superstorms (his “tempestites”). Mylroie (2018) did not ever mention chevron ridges (Hearty, Neumann, and Kaufman, 1998), a major component of the superstorm trilogy. Eleuthera's megaboulders (Hearty, 1997) and their possible transport mechanisms (e.g., Hearty and Tormey, 2018; Rovere et al., 2017, 2018) also are an important matter under discussion.

Initially, we have some remarks about the body of evidence supporting the superstorm trilogy, the published stratigraphy of Eleuthera, and “controversial AAR results” (where AAR is Amino Acid Racemization Geochrology). Subsequently, we respond to Mylroie's issues related to evidence supporting superstorms, “cliff-top, roll-down,” and other challenges.

REGARDING THE BODY OF EVIDENCE

Over several decades, when combined with the works of Kindler and other colleagues (e.g., Godefroid and Kindler, 2015; Godefroid, Kindler, and Nawratil de Bono, 2010; Kindler et al., 2011; Viret, 2008; see also Table 1 herein), along with numerous other colleagues, we have published over 25 peer-reviewed papers on BTC islands (Table 1), logging over 400 sites, and analyzing well over 2000 petrographic and AAR samples. The research database for BTC is substantial, and we encourage readers to examine our previous relevant contributions (Table 1; see also Hearty and Tormey, 2017, Supplement), which are largely uncited here.

Table 1. 

Summary of peer-reviewed publications of regional scope related to issues raised in Mylroie (2018).

REGARDING NORTH ELEUTHERA STRATIGRAPHY

Mylroie (2018) suggested that Carew and Mylroie (1995, 1997) presented a “model for the Bahamas” (based largely on San Salvador Island) that is said to include Eleuthera. Despite our numerous published modifications (Table 1), their model has remained unchanged for decades. Most relevant here, the entire middle Pleistocene section is schematically represented in their model by one cross-bedded unit (Figure 1A, Owl's Hole Formation). They asserted (Carew and Mylroie, 1995, p. 17) that “we have identified rocks definitely assigned to this unit on Eleuthera…,” yet no geologic data were provided. We have documented at least three to four full interglacial sequences (MIS 7–13?) in the middle Pleistocene sequence on Eleuthera, each with numerous facies and pedogenic subdivisions (Figure 1B–E; Tables 1–2). Given this high-resolution stratigraphy, these rocks serve as a premier type section of the middle Pleistocene (Hearty, 1998, ff.). An undifferentiated Owl's Hole Formation does not suffice as representative of middle Pleistocene stratigraphy in BTC.

Figure 1. 

Illustrations of stratigraphic columns and other features related to this reply: (A) The stratigraphic column of Carew and Mylroie (1995) and Mylroie (2008) for comparison with the following. (B–E) Stratigraphic cross sections and photos of predominantly mid-Pleistocene strata just north of Glass Window Bridge (Figure 1B, D; Hearty, 1997, 1998) compared with a schematic section (E) in the same area by Godefroid, Kindler, and Nawratil de Bono (2010). (C) Middle Pleistocene strata (Geology cover photo, Hearty et al., 1999, 27[4]) at Goulding Cay Quarry just 1.2 km SE of the boulders. (F) Holocene and recently transported boulders 1 km NW of the megaboulders showing abundant voids beneath them. Cave formation per se is not required to explain the voids beneath the megaboulders. (G) Speleothems formed in a laterally extensive MIS 5e wave-cut notch at ∼+15 m on the tectonic island of Curaçao. (Color for this figure is available in the online version of this paper.)

Table 2. 

Stratigraphic units defined by Kindler and Hearty (1996, 1997).

We have also addressed the morphological evolution of Eleuthera during the last interglacial (Hearty and Tormey, 2017, Supplement), as wave attenuation on variable-morphology coastlines is a key principle of our superstorm model. In Hearty (1998, p. 336), we observed: “Large fractures along the cliffs suggest…that the bank margin retreats by spalling enormous blocks of limestone…” and that “significant cliff retreat occurred after the emplacement of the last interglacial sequence.”

REGARDING THE RELIABILITY OF AMINOSTRATIGRAPHY

No geochronologic method is perfect, and many have criticized AAR as “controversial,” perhaps appropriately in some past cases. This subjective view was most recently reiterated in Rovere et al. (2017) and Mylroie (2018), yet it is totally unsupported by the facts. Major stratigraphic units (0–VIII) were defined in the BTC by Kindler and Hearty (1996, 1997) on the basis of superposition and petrology (Table 2). In total, 275 whole-rock (WR) samples were collected and analyzed across the BTC. The AAR data demonstrate stratigraphic order in most cases, except in samples from the basal Unit 0, which is correlated with >MIS 13 (∼0.5 Ma). Many unreliable Unit 0 values are published again here (Table 3) in order to explain the important age limitations of the AAR method (e.g., Hearty and Kaufman, 2000; Hearty and O'Leary, 2008).

Table 3. 

Summary of whole-rock A/I values from MIS 5e, 7, 9/11, and >13 rocks (post-5e data not included; stratigraphy trumps AAR) from Eleuthera Island. Aggregate mean values are provided for stratigraphic units. The basal units from The Cliffs, Boiling Hole, and Goulding Cay Quarry are stratigraphically the oldest unit in Eleuthera. The amino acid concentration from this Unit 0 (Kindler and Hearty, 1996) is generally too low (reflecting great age, diagenesis, and recrystallization) to yield reliable A/I values. Samples in bold indicate samples beneath the megaboulders (MIS 5e), or from the megaboulders themselves (MIS 9/11 or >13). Samples were prepared by P.J. Hearty and analyzed at the Amino Acid Geochronology Laboratory in Flagstaff, Arizona (D. Kaufman, Director).

From Eleuthera (Table 3), Unit I, correlated with MIS 9/11 (oolitic/peloidal limestone), yields an average ratio of alloisoleucine to isoleucine (A/I) of 0.647 ± 0.061 (N = 13) from in situ rocks. A/I values from the megaboulders average 0.672 ± 0.062 (N = 5) and clearly correlate with Unit I, except Boulder #4 from Unit 0. The oolite beneath two of the megaboulders (Cow #2 and Hole-in-Rock #4) yielded A/I values of 0.400 ± 0.003 and 0.385 ± 0.004, respectively, which correlate precisely with the Units IV/V island mean of 0.364 ± 0.044. Eleuthera data correlate precisely with the regional (non-random) mean of the MIS 5e oolite A/I ratio of 0.40 ± 0.03 (N = 90; 75 sites including chevron and runup samples from 13 island groups; Hearty and Kaufman, 2000, their Figure 3).

Thus, in BTC, AAR stratigraphic superposition is maintained in over 85–90% of samples (Tables 1 and 3). In regard to our data, we challenge geochronologists (¹⁴C, U/Th, ⁸⁷Sr/⁸⁶Sr, optically stimulated luminescence) and AAR critics to stand up to the same rigorous tests of rock-stratigraphic superposition. Given our large database, unbiased scientists should consider this AAR data to be a reliable measure of relative stratigraphic age of the deposits, particularly given that over 95% of deposits in BTC do not contain corals.

REGARDING “RAIN-INDUCED FENESTRAE”

“Bubble” fenestrae typically define an intertidal (beach) environment (e.g., Dunham, 1970; Shinn, 1983); however, they commonly occur high in MIS 5e chevron and eolian ridges across the BTC (Hearty, Neumann, and Kaufman, 1998; Hearty and Tormey, 2017; Wanless and Dravis, 1989). To explain dune fenestrae, Bain and Kindler (1994) proposed a rainfall origin based on two localities: Annie Bight, Eleuthera (+43 m), and Watling's Castle, San Salvador (+12 m). Tormey (1999, pp. 72–74) contested the rainfall model, showing in detailed outcrop maps (Figure 2; Tormey, 1999, plates 1–5) that fenestral porosity is concentrated in the low-angle backsets and topsets of the eolianites, not in the high-angle foresets or interdune swale. Tormey (1999) further demonstrated that the character and abundance of fenestrae in lowland chevrons (Hearty and Tormey, 2017, their Figure 10AB) were noticeably different from those of high-elevation eolianites (Hearty and Tormey, 2017, Figure 10C), a critical distinction that Mylroie (2018) apparently missed.

Figure 2. 

(A–D) A single 1 m² of outcrop at Annie Bight, Eleuthera (+43 m), described and sampled by Tormey (1999). The location is 49.7 m from NE end of the road cut, in the south highwall, 1.3 m above the road. Tormey (1999) reported six distinct fenestral beds in this 1 m², up to 3 cm thick. The 1 m² was photographed (A) in July 1999, and again (B) in October 2014. The nails, string (white arrows), and paint marking the 1 m² PVC frame in 1999 were still present at all the sites in 2014. (C) The sampled fenestral horizon can be traced laterally for 6 m, with a tabular dip of <2° seaward to the NE. (D) Thin section confirms equant fenestrae. (E) Detailed outcrop map of the road cut at Two Pines, Eleuthera (+26 m). Note the fenestral horizon (dots) beneath the recognized protosol, as well as fenestral horizons above. (Color for this figure is available in the online version of this paper.)

Kindler and Strasser (2000) argued that associated eolian bedding structures support a strictly eolian origin for supratidal fenestrae. Hearty, Tormey, and Neumann (2002) countered that a mix of wave and eolian structures is typical of supratidal environments, particularly during and after storms, and that MIS 5e chevrons contain both waterlain fenestral bedding and eolian structures (Hearty, Neumann, and Kaufman, 1998; Tormey, 1999).

Mylroie (2018) stated, “There is no convincing evidence presented by Hearty and Tormey (2017) to disprove the rainfall slurry.” We ask, how does one provide evidence to disprove something that is virtually unknown in the geologic record? Proponents of the rainfall slurry model are encouraged to publish in reputable journals independent stratigraphic descriptions, and supporting first-hand documentation (in a rainstorm) of this elusive process. If seasonal rainfall creates fenestrae in dunes, it should be universally observed in carbonate eolianites of all ages.

REGARDING THE CLOSE INSPECTION OF OUTCROPS

Outcrop maps precisely locating sedimentary structures (e.g., fenestrae, rip-up clasts, scour, protosols, roots) in MIS 5e chevrons and eolianites have been produced from several islands (e.g., Hearty, Neumann, and Kaufman, 1998; Tormey, 1999; Tormey and Donovan, 2015). Tormey (1999) also collected comprehensive statistical data on primary structures (e.g., abundance and orientation of laminae and fenestral beds in a 1 m² polyvinyl chloride frame; repeated 36 times) at 10 localities on Eleuthera (Figure 2A–D).

Mylroie (2018, his Figure 2) asserted that we misinterpreted clasts within a protosol at Two Pines, Eleuthera, as “storm surge rip-ups” (Mylroie, 2016), and “the result of washover scour” (Mylroie, 2018). This assertion is simply not true. Tormey (1999, pp. 38–40) thoroughly described this “protosol” (as defined by Vacher and Hearty, 1989) at Two Pines. There is no association of the protosol with runup; rather the latter features occur lower in the section (Figure 2E). Mylroie (in Kindler et al., 2010, p. 29) also reported the “occurrence of fenestral porosity in the lower unit suggest[ing] the possibility of marine washover[sic]” at Two Pines.

We have documented rip-up and scour at multiple localities (Hearty and Tormey, 2017, Figures 11–12), in some cases with scour structures indicating uphill flow at +23 m (Hearty, Neumann, and Kaufman, 1998, Figures 7–8). Despite this, Mylroie (2016) stated: “…the dunes at all elevations show no evidence of scour.” As Tormey (1999) demonstrated, finding the evidence requires close, detailed inspection of each outcrop. Rather than dismissing the evidence as “myth” (Mylroie, 2016), might it be that some are not looking closely enough?

REGARDING THE MEGABOULDERS

Boulder names, numbers, and descriptions (Bull #1, Cow #2, Maverick #3 [mid island], Hole in Rock #4] 0.3 km, Soundside shore], and Twin Sisters #5 [0.5 km in Exuma Sound]) have been published in Hearty and Kaufman (2000, Appendix A) and Hearty and Tormey (2017, Supplement, p. 7). Since 1997, we have described the “local phenomena” surrounding the megaboulders, including the unique, deep horseshoe bathymetry and paleogeomorphology. Mylroie (2018, p. ##) stated, “the boulders were far bigger when they were emplaced…,” and we agree, having noted in previous publications the: “…high erosion and dissolution rates on the exposed surfaces” (Hearty, 1997, p. 332), and “increased density by cementation, and decreased volume by dissolution of the rocks…over the past 120 kyr” (Hearty and Tormey, 2017, p. 354).

REGARDING A SUBBOULDER PALEOSOL

There is no terra rossa soil between the megaboulders and underlying MIS 5e oolite. Hearty (1997, p. 332) stated: “The pedogenic calcrete encircles the base of the boulders, marking the previous level of the land surface, and is not developed on the strata beneath them” (Hearty, 1997, Figure 7C, p. 334).

The voids beneath the boulders contain speleothems, which necessarily must have formed after the boulders were deposited. Accordingly, U/Th ages confirm younger, post–MIS 5e ages. The earliest flowstone (2–3 mm) that developed on the MIS 5e oolite beneath the largest boulder Bull #1 produced a U/Th age of 109 kyr. A 46-mm-thick flowstone beneath boulder Hole-in-Rock #4 produced U/Th ages 90–64 kyr (Hearty and Tormey, 2017, Figure 8). Viret (2008) analyzed two additional samples from one stalactite formed on the base of Boulder #2, which yielded ages of 82 and 56 kyr, respectively.

REGARDING SUBBOULDER “CAVES” AND SPELEOTHEMS

Mylroie (2018) faulted us for not addressing a “cave issue,” maintaining that caves were formed at the base of the boulders by a “freshwater body…at least 2 m thick.” We disagree with both the cave formation and groundwater model. Many Holocene megaclasts have voids beneath them (Figure 1F). Irregular surfaces may have formed before the rocks were transported. Other possible mechanisms are fracturing during transport, post-transport mechanical erosion of the softer sediments beneath them, and/or dissolutional processes, including the focus of rain and seawater by the flagging of tall boulder structures.

Speleothems beneath the boulders also are not necessarily diagnostic of “caves” (Figure 1G). On Cayman Brac, Jones (2010, p. 15) described the formation of speleothems in MIS 5e wave-cut notches, stating, “Speleothems…are typically associated with mineral precipitation in the dark, enclosed environments of caves. Such precipitates, however, are not confined to caves because they can grow wherever water drips from bedrock into an open cavity or from an overhanging ledge.” U/Th dating showed the speleothem growth coincided with the MIS 5a to 4 and MIS 2 to 1 transitions, long after the notch was cut during MIS 5e. (Jones, Zheng and Li, 2018).

REGARDING KARST TOWERS

Mylroie (2018) stated, “The tower karst model was presented as a possible alternative to a bank margin collapse, tsunami, or storm origin for the boulders…but the tower karst model has not been disproved” and he asks “…where are all the other [transported mega-] boulders?” Our initial reaction: Where are all the other karst towers or “karrentisch?”

The formation of tower karst as modeled by Mylroie (2008, 2018) requires a specific series of major erosional and dissolutional events to fortuitously occur in a complex paleotopographic setting of high-relief dune and swale, all traces of which apparently have completely disappeared, leaving only the megaboulders. The implication is that this process, along with 2 m of fresh groundwater, occurred at each of the seven megaboulders in their own unique settings and topography, and nowhere else in BTC. The tower karst hypothesis could be easily tested by stratigraphy and petrography, but it is already disproven by the published stratigraphy of the cliffs strata (Figure 1B–D). Finally, the karst tower idea is directly contradicted by both of Mylroie's (2018) alternatives of roll down and margin failure.

REGARDING ROLLING AND SLIDING OF MEGABOULDERS

As in Table 1, Hearty and Tormey (2017, 2018) have repeatedly maintained that the boulder site cliff stratigraphy does not accommodate the “cliff-top, roll-down” hypothesis of Rovere et al. (2017, 2018). Despite this, Mylroie (2018, p. xx) persisted: “Wave energy need not be enough to lift blocks up and over a sea cliff, but only enough to fracture the cliff edge such that the boulders roll or slide downslope landward onto the younger, on-lapping eolianites to create both the age inversion and dip of beds reported by Hearty and Tormey (2017).”

Mylroie's (2018) model (his Figure 6) from The Cliffs (22 km SE) is misleading. In reality, the average seaward slope of the outcropping middle Pleistocene cliffs around Cow and Bull is considerably less (see Figures 1B–D), making it just as likely that the paleocoast sloped downward hundreds to ∼1000 m further seaward.

We challenge the “cliff-top, roll-down” hypotheses for the following reasons:

REGARDING A CATASTROPHIC END OF MIS 5e

Our superstorm trilogy provides a snapshot of the past impacts of climate change in the Atlantic Ocean at the close of MIS 5e. Our hypotheses are supported by dozens of sites and hundreds of analyses in multiple disciplines, published in numerous peer-reviewed papers (Table 1). Our trilogy demonstrates the differential effects of large waves crashing onto the coastal geomorphology of the BTC. Giant waves are not likely attributed to local hurricane events, but more likely super versions of the “Perfect Storm” (late October 1991) in the North Atlantic Ocean. We recognize there may be alternative hypotheses, but at this time, the lion's share of field evidence points to superstorms impacting coastlines across the BTC at the end of MIS 5e.

Our current climate is being dramatically altered by the extraordinary events of the Anthropocene: We can observe and document rapidly rising CO₂ levels, global air and ocean temperatures, and acidification of the oceans; disappearing Arctic summer sea ice; increasing frequency of extreme weather events; and rising sea levels. Extreme caution is advised as our studies of past natural cycles indicate ominous signs of potential severe effects of climate change in the future.