Bolide impact effects on the West Florida Platform, Gulf of Mexico: End Cretaceous and late Eocene

The West Florida Platform constitutes the western half of the Florida Platform, a large sedimentary edifice (dominantly Mesozoic and Cenozoic carbonates) that separates the eastern Gulf of Mexico from the western North Atlantic Ocean (Fig. 1). The study area extends ~700 km from DeSoto Canyon in the northwest to the Straits of Florida in the southeast and averages ~300 km in width from the Florida Escarpment to the Florida shoreline; total area is ~200,000 km² (Fig. 2). The general stratigraphic framework of the West Florida Platform is well known (Salvador, 1991a; Randazzo, 1997; Snedden and Galloway, 2019), and several detailed studies focused on limited geographic areas and/or stratigraphic intervals (e.g., Dobson, 1990; Dobson and Buffler, 1997; Jee, 1993). Seismic reflection data analyzed in this study indicate that two unusual disruptions of normal depositional processes on the West Florida Platform can be correlated with bolide impacts: one at the end of the Cretaceous Period (Chicxulub event, impact location in Mexico); the other during the late Eocene Epoch (Chesapeake Bay event, impact location beneath Virginia, USA, portion of the Chesapeake Bay).

For this investigation, I analyzed migrated, two-dimensional, multichannel seismic reflection data from seven different survey programs, which are publicly available online from the U.S. National Archive of Marine Seismic Surveys (https://walrus.wr.usgs.gov/NAMSS; Fig. 3; Table 1). Because different companies collected the seismic data in different years (1974–1988) using different surveying and processing systems, seismic resolution varies among the data sets. In order to provide more detailed interpretations and illustrate them more clearly for the reader, I have compressed the seismic reflection lines horizontally and stretched them vertically, which exaggerates the architecture of the features (vertical exaggeration roughly 11:1–17:1 at the seafloor). A total of 327 lines yielded 40,405 line km of data (Table 1). In addition, I acquired biostratigraphic data—mainly last occurrences of key planktonic foraminifera and calcareous nannofossils—from 28 petroleum-industry boreholes. Twenty-four are located in federal waters and four in state waters (Fig. 3; Table 2). Borehole data from federal waters may be obtained online from the U.S. Bureau of Ocean Energy Management (https://boem.gov). I personally analyzed foraminifera and ostracoda from the Texaco 2523-1 borehole but relied on public data (both published and unpublished) for the remaining boreholes. All stratigraphic levels identified in industry boreholes are approximate because samples were taken at meter-scale intervals and all samples were cuttings, not cores. Several shallow coreholes (Mitchum, 1978; Mullins et al., 1988a) and seafloor dredges (Freeman-Lynde, 1983; Mullins et al. 1988b; Paull et al., 1990) as well as four short shelf-edge seismic lines (Corso et al., 1988) provided additional data from the outer platform slope and escarpment face (Fig. 4).

Early seismic profiling using the arcer method, along with seismic refraction and reflection studies in the 1960s, established the general physiography, structure, and stratigraphic succession of the northwestern half of the West Florida Platform (Antoine and Harding, 1965; Antoine and Jones, 1967; Antoine et al., 1967; Uchupi and Emery, 1968; Bryant et al., 1969). These studies along with at least one piston core from the escarpment documented presence of a linear shallow-water carbonate buildup (“reef”) along the Early Cretaceous platform edge. Mitchum (1978) used seismostratigraphic sequence analysis (sparker and shallow borehole data) to document a major shift on the outer West Florida Platform from shallow-water shelf deposition in the Early Cretaceous to deep-water slope conditions in the Cenozoic. Ball et al. (1988) published the earliest analysis of a comprehensive grid of multifold seismic reflection lines (1770 line km) tied to 17 deep industry boreholes and interpreted basement structures and general stratigraphy of the northwestern and central parts of the West Florida Platform. Salvador (1991a) published a geological synthesis of the entire Gulf of Mexico basin. Summary chapters in this compendium included specific stratigraphic intervals on the West Florida Platform: Triassic–Jurassic (Salvador, 1991b); Lower Cretaceous (McFarlan and Mendes, 1991); Upper Cretaceous (Sohl et al., 1991); and Cenozoic (Galloway et al., 1991). Snedden and Galloway (2019) published the most recent basin-wide analysis of depositional and paleogeographic history, updating earlier interpretations with abundant new seismic and borehole data. Three of their schematic dip lines cross the Florida Escarpment and adjacent West Florida Platform.

The most recent regional studies of the West Florida Platform are unpublished theses of Dobson (1990) and Jee (1993), both of which covered its central and northwestern segments and emphasized seismostratigraphic sequence analysis. Dobson (1990) investigated pre-Jurassic and Jurassic sections and later published the principal results (Dobson and Buffler, 1991, 1997). Jee (1993) focused on Upper Cretaceous and Paleogene intervals, with emphasis on Eocene rocks; a few of his results were published by Randazzo (1997).

Widespread karstification of Cenozoic strata precludes high-quality seismic reflection data from Mesozoic strata in the southeastern segment of the West Florida Platform, and no previous comprehensive studies of that segment have been published. However, several authors have published more localized studies in that area (Shaub, 1984; Macurda, 1988; Denny et al., 1994).

A wide variety of additional studies focused on specific geographic areas and stratigraphic intervals of the West Florida Platform as well as adjacent coastal Florida. Cretaceous studies include Winston (1971, 1978), Applegate et al. (1982), Freeman-Lynde (1983), Applegate and Pontigo (1984), Applegate (1987), Corso (1987), Corso et al. (1988, 1989), Mullins et al. (1988b), Faust (1990), Paull et al. (1990, 1991), Gardulski et al. (1991), Hine (1997), and Randazzo (1997). Previous reports related to Chicxulub impact deposits in the eastern Gulf of Mexico include Dohmen (2002), Denne et al. (2013), Sanford et al. (2016), and Poag (2017). No specific studies of the Eocene section of the West Florida Platform have been published.

Previous authors generally described the West Florida Platform as having been a rimmed carbonate shelf during most of Early Cretaceous time (Bryant et al., 1969; Corso, 1987; Hine, 1997; Randazzo, 1997; Hine and Locker, 2011; Snedden and Galloway, 2019). Mullins et al. (1988a, 1988b) described the outer part of the modern shelf as a ramp slope, which began to develop in the late Oligocene. At its widest point, the modern “shelf” segment of the West Florida Platform slopes gently southwestward from the Florida shoreline to ~250 km offshore (Fig. 5). There (~200 m water depth), a notable increase in seafloor inclination marks the inner edge of the ramp slope. Seafloor inclination increases dramatically ~300 km from shore (~1.5–2 km water depth) at the lip of the escarpment as it plunges to ~2–3 km below sea level. These aspects, especially shelf and slope width, seafloor morphology, escarpment height, and roughness of the escarpment face, exhibit moderate to notable variability along depositional strike (e.g., Twichell et al., 1990; Figs. 6–12). Locker and Buffler (1983) reported a range of ~15° to ~40° for the escarpment slope angle at three locations near 27°00′N, whereas Paull et al. (1991) reported roughly 60° slope angles at three locations near 25°00′N. Using vertical exaggeration of generally 13:1, my data from 12 locations along the escarpment indicate slope angles of ~75°–80°; with no vertical exaggeration, slope angles range from ~15° to 55° (Table 3). The escarpment face is only moderately rough along its northwestern segment, cut by numerous small ravines (Twichell et al., 1990; Fig. 6A), but the southeastern segment is steep, jointed, terraced, and notched by numerous deep box canyons and shallow gullies (Fig. 6B). Canyon walls are terraced as well and are cut, in turn, by multiple series of linear gullies (Fig. 7; Paull et al., 1990, 1991). Thick Cenozoic bathyal clastics bury the escarpment base (Paull et al., 1990, 1991; Galloway et al., 1991) and bury the entire escarpment at its northwestern and southeastern extremities (Fig. 8A).

Lower Cretaceous rocks and underlying Upper Jurassic strata constitute, by far, the thickest depositional units of the West Florida Platform, and the Lower Cretaceous section dominates outcrops on the Florida Escarpment. The Lower Cretaceous section reaches as thick as 4000 m in boreholes of the Apalachicola Basin (e.g., borehole Sohio 3890-1; Fig. 3; Fig. S1¹) but thins to ~900 m in nearshore boreholes (Table 4). Seismic data indicate similar thick sections in undrilled downdip regions. For example, Dobson and Buffler (1991) estimated ~4000 m of Lower Cretaceous section at the downdip margin of the Tampa Embayment approximately along seismic line GWT86-1 (Fig. 3; Fig. S2 [^{footnote 1}]). Despite the major importance of Lower Cretaceous rocks in the stratigraphic and paleoenvironmental development of the West Florida Platform, few detailed regional studies address these rocks, though several important more localized studies are available (e.g., Addy and Buffler, 1984; Applegate, 1987; Corso et al., 1988, 1989; Locker and Buffler, 1983; MacRae and Watkins, 1992; Randazzo, 1997). McFarlan and Menes (1991) and Snedden and Galloway (2019) included generalized regional reviews of this stratigraphic interval for the entire Gulf of Mexico.

Depending on the type of data applied (microfossil and sedimentological samples versus seismic reflection data), interpretations of end-Cretaceous depositional disruption have reached opposing conclusions regarding escarpment origin and paleoenvironments.

Researchers have collected 119 scattered samples of litho- and biofacies from the Florida Escarpment. Because most samples lack evidence of platform-margin paleoenvironments, Freeman-Lynde (1983) and Paull et al. (1990) concluded that much of the original margin of the platform must have been eroded shoreward, possibly by ~5–10 km. However, only six of these 119 samples contained Early Cretaceous (Aptian, Albian) foraminifera and nannofossils (Antoine et al., 1967; Bryant et al., 1969; Freeman-Lynde, 1983; Paull et al., 1990). Furthermore, none contained Jurassic microfossils, though seismic data indicate that Jurassic rocks crop out at several places along the escarpment face. Six Lower Cretaceous samples from the roughly 900 km² exposed on this escarpment are clearly not representative of the entire outcrop. Much more sampling is required for a conclusive paleoenvironmental interpretation.

In contrast to Freeman-Lynde (1983) and Paull et al. (1990), Locker and Buffler (1983) interpreted the escarpment's seismic signature to represent a platform-edge reef barrier. Corso et al. (1988) agreed that chaotic seismic reflections and diffraction hyperbolae at two locations across the escarpment edge indicated the presence of reefal platform-margin facies. Corso et al. (1988) also concluded that the original platform edge had eroded shoreward ~5–10 km during a ~40 m.y. interval from mid-Cenomanian through late Paleocene time. These authors used a geometric method to derive the amount of retreat. They estimated the distance between the base of the modern escarpment and the point at which a prominent couplet of seismic reflections truncates the top of the Early Cretaceous “toe-of-slope” facies (Fig. 13). They correlated this seismic couplet with a “mid-Cretaceous unconformity” or “mid-Cretaceous sequence boundary,” which was widely recognized across the deep Gulf of Mexico basin. However, later studies (Dohmen, 2002; Denne et al., 2013; Sanford et al., 2016; Poag, 2017) have shown conclusively that the so-called “mid-Cretaceous unconformity” represents the final pelagic facies of an enormous gulf-wide body of sediment (>198 × 10³ km³; Sanford et al., 2016) derived from widespread effects of the Chicxulub bolide impact. Thus, the Early Cretaceous “toe-of-slope” facies of Corso et al. (1988) and its capping seismic couplet compose part of what many researchers have called the Cretaceous-Paleogene boundary deposit (e.g., Sanford et al., 2016; Snedden and Galloway, 2019). This terminology is misleading because the International Commission on Stratigraphy (Molina et al., 2006) has redefined the Cretaceous-Paleogene boundary and assigned all deposits resulting from the Cretaceous-ending impact to the earliest Paleocene. Because the chronostratigraphic terminology applied to this unit is potentially subject to future change, I prefer to tie it to a specific geologic event and call it the Chicxulub impact deposit (Fig. 13). Though the Chicxulub impact deposit in the Gulf of Mexico contains abundant Cretaceous-age debris, that debris was redistributed and deposited during the earliest Paleocene. This means that the false “mid-Cretaceous unconformity” of the basin is not equivalent to the genuine mid-Cretaceous unconformity of the escarpment. These relationships mandate modification of the Corso et al. (1988) method of measuring escarpment-retreat distance.

In light of this new interpretation, I propose a method to estimate “minimum escarpment retreat” (MER). It is similar to the method of Corso et al. (1988) but measures the lateral distance between the modern location of the platform margin (i.e., the top of the Lower Cretaceous escarpment) and the estimated position of the base of the escarpment prior to deposition of the Chicxulub impact deposit (roughly the base of the preserved Jurassic platform margin; Fig. 13).

The estimate is minimum because, presumably, part of the Jurassic platform margin would have been removed as well. However, the platform-edge position of the Upper Jurassic section is not always clearly evident. My measurements indicate average MER of ~7 km, ranging from ~3 to ~10 km (Table 5). Greatest MER (~10 km) is indicated for seismic lines Dv88-33 and GWT86-16 (Fig. S3 [^{footnote 1}]). All MER values are approximate, however, because seismic data are partly obscured by chaotic reflections (hyperbolic diffractions, velocity pullups) on most cited lines. Dillon et al. (1988) applied synthetic seismograms to demonstrate, for example, that the actual location of an escarpment face corresponds approximately to the apices of the hyperbolic diffractions, significantly shoreward (by several kilometers) of its apparent location on the seismic reflection image (Fig. 14).

Most authors (e.g., Corso et al., 1988; Paull et al., 1990) have concluded that shoreward retreat of the platform edge took place through relatively slow-acting biological, chemical, and gravitational processes of erosion. However, recent documentation of the far-reaching effects of the Chicxulub bolide impact give reason to consider a more rapid retreat process. For example, Paull et al. (2014) concluded that massive margin collapse of the nearby Campeche Escarpment must have contributed significantly to the broad breccia-rich sediment apron flanking the escarpment and suggested that a collapse of the Florida Escarpment produced a similar sediment apron. These authors documented the presence of large megablocks (~2 km wide) within the apron deposits, inferring their original source to have been the escarpment face. Two seismic reflection profiles across the northwestern flank of the Campeche Platform published by Locker and Buffler (1983) imaged the Chicxulub impact deposit there (Figs. 15, 16). Line GT3-60 (Fig. 16) documents large slump blocks (~10 km in length) in the impact deposit, expressing parallel, stratified seismic facies. They appear to have been derived from the Late Cretaceous escarpment face.

Numerical modeling indicates that seismic shaking and ground roll from Chicxulub would have produced >1 m of vertical motion on the West Florida Platform within five minutes of impact, followed in ~1 h by tsunami-driven erosion and sediment redistribution (Sanford et al., 2016). These mechanisms severely disrupted Upper Cretaceous strata over ~64 × 10³ km² of the West Florida Platform, reaching as far as 180 km shoreward of the escarpment (Figs. 17A, 17B; Poag, 2017). A potential sediment volume of ~4–5 × 10³ km³ was redistributed from the platform interior to the Chicxulub impact deposit apron bordering the escarpment. These circumstances imply that a significant amount of platform-edge erosion must also have incorporated instantaneous, impact-related processes operating in the final moments of the Cretaceous. The resultant debris, represented in part by chaotic, discontinuous, high-amplitude seismic reflections in the Chicxulub impact deposit (“toe-of-slope” facies), accumulated in the earliest moments of the first day of the Paleocene (Gulick et al., 2019). The Chicxulub impact site's relatively close proximity (~600 km) to the southeastern end of the West Florida Platform (Fig. 18A) would account for the extensive jointing and canyon cutting along this southeastern part of the escarpment. Florida Canyon, the deepest and longest escarpment canyon, is located there (Fig. 18B).

Eocene strata are well documented on the Florida peninsula (Chen, 1965; Miller, 1986; Randazzo, 1997) and are dominantly inner to middle neritic carbonates. In contrast, the only specific study of Eocene strata of the West Florida Platform is the unpublished Ph.D. thesis of Jee (1993). However, Mitchum (1978) and Mullins et al. (1988b) included analyses of Eocene strata in core samples derived from the outer part of the Eocene ramp slope, which consisted mainly of pelagic deposits dominated by foraminifera-nannofossil ooze or chalk (Gardulski et al., 1991). On the other hand, the Eocene section updip on the shelf, such as in the Texaco 2523-1 borehole, contains mainly dolomitic inner neritic deposits dominated by benthic foraminifera (Fig. 19). The paleophysiography of the Eocene section differs strongly from that of the Lower Cretaceous section in that the shelf edge is not located at the escarpment but as much as 100 km or more updip of the escarpment (Figs. 20, 21; Fig. S4 [^{footnote 1}]). The Eocene section is roughly 800–1000 m thick in boreholes on the shelf (Table 6) and thins significantly basinward across a broad slope ramp to 70–120 m in shallow coreholes located near the top of the escarpment (Mitchum, 1978).

A notable characteristic of the upper Eocene section on the West Florida Platform is a relatively narrow zone (~11–28 km wide) of faulted and/or mounded sediments that extends for ~270 km along the upper edge of the Florida Escarpment (Figs. 21–25). Randazzo (1997) illustrated this disrupted unit along a seismic reflection profile that approximates profile GWT86-19. In the context of the Lower Cretaceous shelf, this might be explained as a reef-rimmed margin. However, the pelagic water depths this far basinward from the Eocene shelf edge would have been too deep for reef development.

The seismic data indicate that most of the mounded zone contains two distinct types of disrupted strata. Where both types are present, the lower layer displays high-amplitude, continuous reflections, broken up into distinctive en echelon block faults (Figs. 22, 23; Figs. S5–S9 [^{footnote 1}]). In contrast, the upper layer displays low-amplitude, discontinuous, chaotic reflections (folds, fractures, slumps, faults) similar to those typical of the disrupted Upper Cretaceous strata of the West Florida Platform (Fig. 17; Poag, 2017). I interpret this bipartite arrangement to indicate in situ disruption of the lower, high-amplitude layer resulting from seismic shaking and ground roll, whereas the upper chaotic layer probably represents debris displaced from updip locations via debris flows, slumps, and slides. This material would have been redistributed from the late Eocene shelf or upper ramp slope as a result of tsunami-driven currents arising from a marine impact. A few sites clearly exhibit the upper layer of redistributed debris (Figs. 24, 25) but with no obvious en echelon faulting in the lower layer.

I infer that the late Eocene Chesapeake Bay impact (Poag et al., 2004), ~1300 km northeast of the Florida Escarpment (Fig. 26), provided the kinetic energy for this two-stage disruption. Applying an impactor size of ~3 km diameter (Collins and Wünnemann, 2005), the online numerical model of Melosh et al. (2004) predicts that the impactor entered the atmosphere at a velocity of ~17 km/s imparting a kinetic energy burst of ~2 × 10⁶ Mt TNT to the seafloor. Major seismic shaking (Richter scale magnitude ~9) traveled from the Chesapeake Bay impact site to the Florida Escarpment within ~4 min after impact. These approximate predictions would vary depending on the precise size, velocity, and approach angle of the impactor, which are not known at present. The subsequent tsunami would have reached the Florida Escarpment ~2 h after impact. A direct oceanic connection existed between the impact site and the West Florida Platform because the Florida peninsula was submerged during the Eocene (Fig. 26).

Widespread evidence of individual bolide impacts usually is limited to the airborne spread of small ejecta particles, such as microtektites, or an unusual abundance of cosmic elements, such as iridium. An iridium spike and associated ejecta derived from the Chicxulub impact are globally well documented (e.g., Smit, 1999; Goderis et al., 2013). The most thoroughly documented far-field effect from the Chesapeake Bay impact is a broad field of microtektites and associated ejecta (North American tektite strewn field), recognized mainly in the Atlantic Ocean and U.S. Coastal Plain (e.g., Koeberl et al., 1996; Glass et al., 1998; Biren et al., 2019).

The ages and origins of ejecta deposits can be verified by objective geochemical and mineralogical methods, whereas recognition of seismic and tsunami effects depends more on subjective observational interpretation of field data. The extensive database collected from industry and academic seismic surveys in the Gulf of Mexico provides a unique opportunity to observe the detailed stratigraphic and depositional history of this region. As a result, disruptive seismic effects from the Chicxulub impact have been most firmly documented in the Gulf of Mexico (e.g., Locker and Buffler, 1983; Corso et al., 1988; Grajales-Nishimura et al., 2000; Paull et al., 2014; Sanford et al., 2016; Poag, 2017). Related studies outside the Gulf of Mexico include the Caribbean region (e.g., Bralower et al., 1998; Tada et al., 2003), the western North Atlantic (e.g., Klaus et al., 2000; Norris et al., 2000), and the Adriatic region (Korbar et al., 2015). The data presented herein from the West Florida Platform support previous interpretations that the Chicxulub bolide impact produced instantaneous widespread platform-margin collapse, resulting in a significant shoreward retreat of the Florida Escarpment. The resultant sedimentary debris collected as a continuous apron along the base of the escarpment. My interpretation recognizes this debris as the Chicxulub impact deposit, which is not a transitional unit between the Cretaceous and Paleogene Periods but accumulated entirely in the first few days of the Paleocene Epoch (Danian Age).

Only a single previous report of wide-field physical disruption from the Chesapeake Bay bolide has been published. Poag et al. (2004) reported initial recognition of possible seismic and tsunami effects from the Chesapeake Bay impact based on observations from submersible dives on the lower continental slope of New Jersey, ~350 km northeast of the impact site. A broad, linear, cliff-faced outcrop (>10–15 km wide) of highly fractured and brecciated lower and middle Eocene pelagic limestone occupies that location (Figs. 27A–27C). The same area features unusual vertical-walled, flat-bottomed erosional channels (3–5 m wide, 4–13 m deep) with axes trending downslope (Robb et al., 1983; Figs. 27A, 27B, 27D). I interpret these features to represent another example of fracturing and brecciation due to seismic shaking and ground roll followed by unusual channeling created by tsunami-generated bottom currents, a two-step disruption similar to the end-Cretaceous and late Eocene disruption phases indicated on the West Florida Platform.

Among the approximately dozen known submarine craters (Poag et al., 2004), far-field seismic effects have been reported from only two—Chicxulub and Chesapeake Bay. Though many published reports have noted such effects attributable to the Chicxulub impact, this is only the second report of such effects derived from the Chesapeake Bay impact. It is the first report of their recognition in the Gulf of Mexico.

I am grateful to the following persons for their assistance in this research: Matt Arsenault for help attaining seismic data; T.J. Christian for providing samples from the Texaco 2523-1 borehole; Deborah Hutchinson, David Foster, Christian Koeberl, and an unnamed reviewer for helpful reviews of the original manuscript; and especially to the U.S. Geological Survey for supporting my continued research as a Scientist Emeritus. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government.