The origin of carbonate turbidites has been the subject of debate since the introduction of depositional models that are guided by sequence stratigraphy. Sequence stratigraphy has postulated that in siliciclastic depositional systems most sediment is shed into the deep basins during lowstands of sea level (Vail et al., 1977). Whether this principle is applicable to the carbonate depositional system is at the heart of the controversy. Scientists working with carbonates have presented evidence that carbonate depositional systems are 180° out of phase with siliciclastic systems. Carbonate platform rates of production are highest when the platform is flooded. During these times, more sediment is produced than can be accumulated on the platform top, and this excess sediment is exported into the adjacent basins (Supko, 1963; Kier and Pilkey, 1971; Lynts et al., 1973; Hana and Moore, 1979; Hine et al., 1981; Droxler et al., 1983; Boardman and Neuman, 1984; Schlager et al., 1994). The higher sedimentation rate may result in overloading of the slope and gravitational instability, which are major contributing factors in the generation of turbidity currents (Middleton and Hampton, 1976; Crevello and Schlager, 1980). Therefore, it is reasonable to expect a higher frequency of turbidity currents during times of high slope-to-basin sedimentation, which for isolated platforms appears to correlate with relative highstands of sea level (Mullins, 1983).
This highstand shedding of turbidites has been well documented in the Quaternary sections surrounding the Bahamas (Droxler et al., 1983; Droxler and Schlager, 1985; Reijmer et al., 1988, Glaser and Droxler, 1991). A similar pattern of deposition has been interpreted from ancient slope sections surrounding isolated platforms in the Triassic and the Cretaceous (Reijmer et al., 1991; Harris 1994; Vecsei and Sanders, 1997). In ancient deposits, analyses of turbidite shedding have commonly relied on the analysis of grain composition. In these methods, the turbidites containing more platform interior grains were assumed to be shed during sea-level highstand, while turbidites containing grains predominantly from the margin were shed in times of low sea level (Reijmer et al., 1991).
In this study we take advantage of the geometries seen on the seismic sections. Seismic-sequence boundaries that indicate sea-level falls were determined at the platform margin of the Great Bahama Bank using erosional features and onlap patterns (Eberli, Swart, Malone, et al., 1997). On Leg 166, five sites were drilled on the transect of the western flank of the Great Bahama Bank. The sedimentary sections drilled at these sites were correlated to a multichannel, high-resolution seismic section on which 17 third-order seismic-sequence boundaries were traced from the platform top down to the basin (Eberli, Swart, Malone, et al., 1997). The seismic and sedimentary data were used to define highstand and lowstand systems tracts within the sequences.
In this paper we show that turbidites are shed during both sea-level highstand and lowstand. In addition, our data indicate a shift of the depositional location with changing sea level. For example, in the Miocene, sea-level highstands produced a high turbidite frequency on the lower slope (Site 1003), whereas the turbidite deposits at the toe of the slope (Site 1007) were dominantly shed during sea-level lowstand. Our analyses also document a relative small amount of turbidites along the entire transect, which suggests that deposition from turbidity currents is not a dominant process in exporting sediment to the deep-water areas.