The cores recovered along the Bahamas Transect contain the sedimentary record of sea-level changes throughout the Neogene, and the interstitial water geochemistry provided evidence of fluid flow through the platform margin of the Great Bahama Bank.
Sea-Level Objectives
An extremely exciting result of drilling during Leg 166 is the discovery that the sea-level changes throughout the Neogene can be monitored in the facies successions in the slope sediments of Great Bahama Bank. Facies changes correlated with seismic sequence boundaries and corroborated the interpretation that these sequences are controlled by sea-level changes. Calcareous nannofossils and planktonic foraminiferal biostratigraphical datums were found at all sites and an age-depth relationship was established. The Geomagnetic Polarity Time Scale (GPTS) of Berggren et al. (1995) transferred the biohorizons into time, and the cores were correlated with seismic data by using the time/depth conversion from the vertical seismic profile in each hole. This allowed us to date the sequence boundaries precisely, achieving two of our major sea-level objectives: (1) the facies distribution within the carbonate sequences was documented, and (2) the timing of sea-level changes in the Neogene was established. To compare the sedimentary record of sea-level changes with the oxygen isotope proxy, a basinal hole (Site 1006) was drilled. This core far exceeded our expectations, as most conventional nannofossil and planktonic foraminifers were found in the Pleistocene to middle Miocene section. In addition, recovery was very high and the preservation of the foraminifers was good throughout the recovered section. Thus, all the prerequisites are in hand to establish an oxygen isotopic record of the sea-level changes in the same transect. The correlation of the sedimentary and isotopic record will enable us to assess the causal relationship between sea-level changes and the sequence stratigraphic pattern, thereby fulfilling the third sea-level theme objective.
To achieve these goals the following four data sets were crucial.
1. Lithostratigraphic data
Core recovery was sufficient (55.3%) to document the facies successions throughout the cores (Fig. 3). The facies successions contain indications of sea-level changes on two different scales. First, there are high-frequency alternations of layers containing more platform-derived material with layers containing more pelagic sediments. In the Pleistocene and Pliocene periplatform section, these alternations are reflected in the ratio between neritic components and nannofossils, and mineralogically between aragonite and low Mg-calcite. In previous studies these cycles were shown to correlate to orbitally forced high-frequency sea level changes in the Quaternary. We found similar decimeter- to meter-scale high-frequency facies changes in the Miocene sections of all the holes. The layers with more platform derived material are generally lighter in color and are better cemented, whereas the more pelagic layers are darker, greenish, and are less cemented and contain small amounts of clays and silt. Sediments with increased platform input are interpreted to have been deposited during sea-level highstands when the platform was flooded and producing sediment. The sediments with higher siliciclastic input indicate erosion of siliciclastic margins, probably in Cuba and Hispaniola. Because of the color changes and the cementation differences, these cycles are also recorded in color reflectance data and the logs. Formation MicroScanner (FMS) data provided a continuous record of these cycles at two sites, which will be analyzed to determine the frequency of these cycles in the Miocene.
A repetitive pattern of facies succession on the order of tens to several hundreds of meters thick documents the sedimentary record of longer term sea-level changes. These larger-scale patterns are also seen in the seismic images. Similar to the small-scale cycles, changes in the amount of platform-derived material indicate periods of high and low sea level. Platform exposure during low sea level is reflected in reduced sedimentation rates, occasionally leading to hardground formation on the slope. In addition, erosion of the platform margin during these periods is documented by the deposition of coarse-grained packstones and floatstones in pelagic-rich background sediments. Erosional truncation, which leads to hiatuses in the biostratigraphic successions, is also observed in the proximal slope sites. Cycles are best developed during sea-level rises and usually comprise the middle part of the sequences. Redeposition of platform carbonates occurs again in the upper part of the sequences. This pattern is best developed in the Miocene sequences when the platform had a ramp-like morphology. In the Pliocene and Pleistocene, the bulk of the sediments, especially in the more proximal locations, is dominated by thick successions of fine-grained material, but turbidites occur preferentially in the upper parts of the sequences. More mass-gravity flow deposits are found in the distal sites, indicating that the steep upper slopes were bypassed by these flows.
2. Petrophysical and Log Data
To complement the cores and fill in recovery gaps, an extensive logging program was performed in the four deep transect holes. At Site 1003, a dedicated logging hole was drilled when hole conditions prevented the logging of the cored hole. The acquired logs provide detailed information on the sedimentary properties and structure of the strata. Log-to-core correlation permits significant interpretations about variations in sedimentation patterns along the transect sites of GBB from the earliest early Miocene to Holocene. The general compatibility of the discrete data points from shipboard petrophysical measurements with the log data supports the integrity of the data sets. This logging data will help correlate the seismic sequence boundaries to the cores. Over 3 km of Formation MicroScanner data were acquired. These data will allow us to analyze the small-scale cycles observed in the cores and assess their frequency. This analysis has the potential to reveal the frequency of orbitally driven climate and sea-level changes in the Miocene.
For an accurate tie of the cores to the seismic data, vertical seismic profiles with the Well Seismic Tool were acquired in each of the deep holes. These data allowed us to calculate an accurate time/depth conversion of the seismic reflections. These experiments proved to be extremely valuable as neither the integrated log velocity nor the shipboard velocity measurements provided an accurate velocity profile in all the holes. Errors of up to 50 m were obtained using the log or discrete shipboard velocity measurements.
3. Biostratigraphic Data
The most crucial task was to establish the ages of the seismic sequence boundaries and their indicated sea-level changes. Calcareous nannofossils and planktonic foraminiferal biostratigraphic datums were used to establish an age-depth relationship for each of the sites drilled as part of the Bahamas Transect (Sites 1003-1007). Many biohorizons from both micropaleontological groups were identified, establishing a framework of intersite correlation (Fig. 4; Table 1). This framework was transferred into time by using the Geomagnetic Polarity Time Scale (Berggren et al., 1995). This latter step was the most important to obtain absolute ages to fulfill one of the objectives of this leg; testing the age consistency of the sequence boundaries observed in the seismic records and logging data in combination with lithological changes observed in the recovered sedimentary section.
Dating the sediments in upper slope carbonate environments is not always an easy task. At the Leg 166 sites, calcareous microfossils were very rare and difficult to identify in some intervals due to dilution and preservation problems. Fortunately, the boreholes drilled more basinward (Sites 1006 and 1007) contained moderately to well-preserved calcareous microfossils throughout their generally continuous records. Refined biostratigraphies for these holes were established by using the core-catcher samples in the Pleistocene to upper Pliocene sections, and samples from discrete clayey layers in the Miocene sections, which contained moderately preserved calcareous microfossils.
Because Site 1006 is located in the Florida Straits, in a distal position from the platform, it is less affected by diagenetic effects and erosional surfaces associated with sea-level falls than those sites drilled higher on the upper slope. This site, with apparently continuous sedimentation, provides a benchmark for establishing the stratigraphic order of the marker species in this region. Most conventional nannofossil and planktonic foraminiferal biohorizons were found in the Pleistocene to middle Miocene interval at Site 1006. Nannofossil biostratigraphic datums were used exclusively to characterize sedimentation in the Pleistocene section of this site, while the Pliocene and Miocene sections were constrained by both planktonic foraminiferal and nannofossil events.
The depths of planktonic foraminiferal and nannofossil datum levels at Site 1006 were consistent with the relative order of these events in tropical regions except for a few cases. One important mismatch was the first occurrences of Globigerinoides conglobatus and Globorotalia margaritae relative to nannofossil events in the same interval. These datums have been assigned ages of 5.7 Ma (Berggren et al., 1985 ) and 6.4 Ma (Berggren et al., 1995), respectively. Chaisson and Pearson (in press) have subsequently modified the age of Gs. conglobatus to be 6.2 Ma in the tropical Atlantic. In either case, both of these events occur below the top of the small Reticulofenestra spp. interval at Site 1006, dated as 6.5 Ma (Sato et al., 1991). This relationship was also observed at Site 1007. Using the nannofossils only in Site 1006's uppermost Miocene section to establish the age-depth relationship infers that the first occurrences of Gs. conglobatus and Gr. margaritae are 6.8 Ma and 7.0 Ma, respectively, in the Bahamian region. The base of Gs. conglobatus became an important datum level to identify an upper Miocene hiatus in the upper slope Sites 1003 to 1005, and 1007 (see below).
A second apparent problem in the order of the biostratigraphic datums occurred near the middle/upper Miocene boundary. In tropical settings, the first appearance of C. coalitus (11.3 Ma) occurs above the last appearance of Globorotalia mayeri (11.4 Ma; Berggren et al., 1995). At Sites 1003 through 1007, Gr. mayeri is recorded at the same level or above the C. coalitus level. However, it has been shown that the top of the Gr. mayeri interval is slightly younger in subtropical regions (Gulf of Mexico, Jamaica, and North Atlantic Site 563) than observed in the tropics (Miller et al., 1994; Zhang et al., 1993). The position of Gr. mayeri relative to C. coalitus at the Bahamas Transect sites is consistent with this observation.
A third inconsistency in the relative position of the datums involves the first occurrence of C. floridanus (13.2 Ma), with respect to the first occurrence of Globorotalia fohsi fohsi (12.7 Ma). At Site 1003, the base of C. floridanus is above Gr. fohsi fohsi but occurs at the top of a slumped interval.
In contrast to Site 1006, unconformities appear in the upper slope sites. The most prominent of these is in the late Miocene, recognized in all slope sites. This unconformity was recognized by the juxtaposition of Gs. conglobatus (last appearance datum [LAD])-6.8 Ma) with D. hamatus (LAD-8.7 Ma). Also observed at this level was the first appearance of Globorotalia cibaoensis (7.7 Ma) and Globigerinoides extremus (8.1 Ma). At Site 1007, the upper part of the small Reticulofenestra spp. interval (LAD-6.5 Ma) was found above the unconformity, indicating the presence of uppermost Miocene sediments (Messinian). Furthermore, the 16-m interval above the 6.5 Ma level contains poorly preserved nannofossils. Above this zone of uncertainty, the sediments are younger than 4.7 Ma, based on the first appearance of Ceratolithus rugosus. It is unclear whether this is the true first appearance or whether it extends further down. Therefore, the interval of poor preservation represents either an interval of very low sedimentation (<1 cm/k.y.) or contains a hiatus that straddles the Miocene/Pliocene boundary. A lithologic break at the base of this interval of poor preservation indicates the presence of a hiatus. At Sites 1003 and 1005, a similar interval of very poor preservation was also found, but it extended well above the unconformity, preventing any age assignment to this interval.
At Site 1007, three erosional surfaces were observed in the Pliocene-Pleistocene section. The top is a modern erosional surface with the uppermost sediments being older than 0.95 Ma. Within the upper Pliocene, an erosional surface is identified based on the co-occurrence of several planktonic foraminiferal events (2.0 to 2.4 Ma). The lower/upper Pliocene boundary is unconformable, based on the juxtaposition of the first appearance of Globorotalia tosaensis (3.2 Ma) and Globigerina nepenthes (4.2 Ma). These unconformities bound an extensive upper Pliocene drift deposit not observed at the other sites, and they are probably caused by current erosion.
Sites 1003 and 1007 recovered middle and lower Miocene sequences, showing four cycles of alternating sedimentation rates (Fig. 5). At both sites, periods of faster deposition (>5 cm/k.y.) occurred from 11.5 to 13 Ma, 15 to 16.5 Ma, 17.5 to 20.5 Ma, and older than 23 Ma. At Site 1007, slow but continuous pelagic sedimentation (<2 cm/k.y.) occurred from 9.5 to 11.5 Ma, 13 to 15 Ma, 16.5 to 17.5 Ma, and 20.5 to 23 Ma. In contrast, these intervals of reduced sedimentation are represented by hiatuses at Site 1003.
In summary, the overall consistency in the relative position of planktonic foraminiferal and nannofossil datums means that a reliable framework for intersite correlation was established for the Bahamas Transect sites. This framework afforded the opportunity to date the lithologic changes and seismic sequences identified along the transect. Determining the timing of the SSBs and lithologic changes provides the critical information to understand the role of sea-level change and its effect on the development of the Bahamian Platform.
4. Sequence Stratigraphy
Sequence analysis was performed on the seismic data prior to drilling (Fig. 6). Erosional truncations and onlap geometries were used to identify seismic sequence boundaries. These geometrical relationships were best observed farther east along the buried platform margins of Great Bahama Bank. The reflections identified as sequence boundaries were traced to the slope and basinal areas of the Leg 166 Bahamas Transect sites. Tracing the reflection horizons is straightforward for the Pleistocene and the lower and middle Miocene sequences. Uncertainties exist for the upper Miocene and lower Pliocene sequence boundaries because the multiples and diffractions generated by the steep GBB margin could not be completely removed. We identified 17 sequences in the Neogene section, which we labeled a-i and k-q, and the basal seismic sequence boundaries were labeled A-I and K-Q. Most of the sequence boundaries are conformable at the Leg 166 sites; however, truncation is observed in the more proximal sites.
The time/depth conversion allowed us to correlate the sequence boundaries to the cores. All sequence boundaries coincide within seismic resolution of about 10 m, with a facies change indicative of a sea-level change. The exact position of the boundaries is recorded on the log data, which has a better resolution than the seismic data. The age at the correlative depth was calculated using biostratigraphic datums and extrapolating sedimentation rates between datums. Although this procedure carries the uncertainty of both the exact position of the biostratigraphic datums and the seismic resolution, the sequence boundaries showed consistent ages along the seismic reflections (Fig. 7). This result is exciting as it confirms one of the major assumptions of sequence stratigraphy: that seismic reflections are time lines.
The ages of the sea-level changes in the Neogene were estimated from the ages of the sequence boundaries. At the proximal sites, where several of the sequence boundaries are associated with erosion that occasionally is detected by a hiatus, the age of the boundaries were taken from the distal conformable locations. In light of the resolution problem, the ages of the sequence boundaries probably have an error bar of approximately 0.2 m.y (Table 2). Nevertheless, the ages indicate that the seismic sequences along the Bahamas Transect record most of the known major (third-order) sea-level changes in the Neogene. Shore-based log seismic correlation and detailed biostratigraphic analysis will refine the age model for these sea-level changes.
Evidence for Fluid Movement through the Margin of Great Bahama Bank
One of the primary objectives of Leg 166 was to investigate the possibility of fluid-flow processes through the margin of Great Bahama Bank. The association of the fluid chemistry and heat-flow sampling program with the sea-level objectives means that changes in fluid chemistry can not only be examined as a function of age and depth, but also within a sequence stratigraphic framework. This approach is therefore fundamentally different from other investigations of pore-water profiles along carbonate platforms which, in most instances, only drilled single holes.
The clearest evidence for active recharge of fluids through the margin of GBB is derived from the non-steady state profiles of both conservative and non-conservative elements in the upper 100 mbsf pore-water profiles obtained from both the northern (Sites 1003-1007) and southern transects (Sites 1008-1009). A zone was identified (the flushed zone) that is confined to the upper 40 mbsf in which there is an absence of geochemical gradients in both conservative and non-conservative constituents (Figs. 8, 9). The geochemical data in this interval shows essentially no change from bottom seawater concentrations for all of the normally measured cations and anions. The flushed zone at Sites 1006 and 1007, which are situated farther from the platform, is reduced in thickness and exhibits small, but nevertheless significant, increases in Sr2+. A similar upper flushed zone, approximately 40 m thick, also is observed at the sites drilled along the more southerly transect. Most of the sites also showed irregular temperature profiles in this part of the sedimentary column. The absence of geochemical gradients and the irregular temperature profiles support the notion that there is advection of seawater through this portion of the sedimentary column. The small increases in Sr2+ at the more distal sites indicate that fluid flow here is reduced relative to the platform.
Below the flushed zone, there is a sharp change in the concentrations of both conservative and non-conservative elements. The SO42- concentration decreases sharply, whereas alkalinity and Sr2+ increase. The nature of these gradients is not steady state and reflects depression by a downward-advecting fluid. An unexpected finding at all sites was the increase in Cl- concentration with increasing depth. This increase is probably a result of the diffusion of Cl- and Na+ from an underlying brine or evaporite deposits (Fig. 10).
Based on the evidence from the seven sites drilled during Leg 166, there is clear evidence that a mechanism exists, which produces active exchange between the upper 40 m of sediments and the bottom waters. At the present time we do not know the precise mechanism involved in the flushing, only that it exists. The observations are, however, consistent with water being drawn into the platform by the mechanism known as Kohout convection (Kohout, 1967). In this mechanism the temperature difference between the platform interior and the adjacent seaways causes underpressure to develop within the platform, which draws water through the flanks of the platform.