In recent years, Neogene cyclic sedimentary successions have been shown to be paced by insolation variations caused by regular changes in the Earth's orbit and the orientation of the Earth's axis, and have been dated by correlation of the sediment cycles to the calculated insolation variations (Hilgen et al., 1991, 1995; Shackleton et al., 1990). At Site 1006, the sediment alternates between a light yellowish or greenish grey nannofossil ooze (neritic, high sea level) and a light grey (white) nannofossil ooze with foraminifers (pelagic, low sea level). In lithologic Units 3-5, the pelagic part is sometimes topped with a firmground or hardground and in Unit 1, the cycles often start with a basal clay layer (Eberli, Swart, Malone, et al., 1997).
These cycles are observed in most of the downhole logs. Compaction, cementation, and quantity of microfossils all affect porosity, which governs the neutron porosity, density, resistivity, and sonic velocity logs. The balance between carbonate and clays is reflected in the photoelectric factor (PEF) log. Calcite has a PEF of ~5 barn/e -, and clays typically have PEFs of 1.5-3.5 barn/e -. In addition, diamagnetic calcite has a negative magnetic susceptibility, whereas paramagnetic clays have a positive susceptibility. Color reflectance at 700 nm correlates with aragonite content (Eberli, Swart, Malone, et al., 1997). The (spectral) natural gamma-ray log has uranium as its principal source; uranium tends to accumulate at redox boundaries, and mark events such as hardgrounds (Serra, 1984; Williams and Pirmez, 1999), and, therefore, does not show ideal cyclic behavior. Potassium and thorium peaks mark occurrences of clay-rich sediment.
The character of the cycles changes at the boundary between Unit 2 and Unit 3, so the succession is treated in two parts: 180-360 mbsf (Fig. 4) and 360-715 mbsf (Fig. 5). The upper part of the hole (Pleistocene-late Pliocene) is not dealt with here.
In Unit 2 (126-360 mbsf), the resistivity (SFLU) log contains regularly spaced cycles recurring about every 1.4 m (Fig. 4). The same cycles are observed in the sonic velocity log, but are less clearly seen in the density and neutron porosity logs and FMS resistivity images (see figures in Eberli, Swart, Malone, et al., 1997) because the wide borehole in Unit 2 causes degraded contact between tool and borehole wall, which these logs require. Assuming a precessional periodicity of 22 k.y. per cycle, the sedimentation rate is ~6 cm/k.y.: about half that suggested by the biostratigraphy. Cycles in magnetic susceptibility measured on core appear to have a longer wavelength, giving a sedimentation rate of ~12 cm/k.y., in agreement with the biostratigraphy (Fig. 6; Table 4, Table 5, Table 6, Table 7). It is likely that susceptibility is tracking the balance between clay and carbonate downhole. It appears that the distinct cycles in resistivity occur twice per precessional cycle. Resistivity is higher where there is more cementation, typically in the neritic-rich sediment. An initial interpretation of the behavior of the resistivity log is that the carbonate bank is exporting neritic-rich sediment twice per precessionally driven sea-level cycle, maybe indicating both highstand and lowstand shedding. The susceptibility traces variations in clay content, which depends on erosion in the source area (Cuba and Hispaniola), and current strength; magnetic susceptibility is largely independent of the sediment supply from the carbonate platform. It seems unlikely that the biostratigraphic picks over this interval should be far wrong, and their ages are well known, since they are astronomically dated in the Pliocene (Berggren et al., 1995a).
In Units 3-5 (360-715 mbsf), cycles in the resistivity, FMS image, and porosity logs were counted (Fig. 4). Each gave slightly different numbers of cycles, but this can be mostly attributed to the different vertical resolutions of the logs (SFLU resistivity ~45 cm, porosity ~30 cm, FMS image ~0.5cm). The resistive beds in the FMS image are mostly thin relative to the intervening less-resistive sediment, suggesting that they mark a distinct (though as yet unidentified) stage in the carbonate bank's cycle. The logged cycles are paced by precessional forcing, giving sedimentation rates comparable with those derived from the biostratigraphy (Fig. 6).
The reliability of the cycle stratigraphic depth-age line for Site 1006 has four main constraints, the latter three resulting from the first one. First, it is difficult to identify long-term eccentricity-controlled amplitude variations in the precessionally controlled log cycles at Site 1006. From the point of view of dating, this is a major problem because amplitude variation is the key to matching a specific sedimentary cycle to the corresponding insolation cycle (Hilgen et al., 1991, 1995; Shackleton et al., 1990). The porosity-related log cycles at Site 1006 are depositional features, but they are also diagenetically enhanced. During both of these processes, some of the proportionality to the original insolation/sea-level forcing may have been lost. Second, the cyclostratigraphy is 'floating,' and needs to be tied to a particular reliable biostratigraphic date. Third, a duration of 22 k.y. for each cycle was chosen (somewhat arbitrarily) to reflect the relative importance of the 23 k.y. over the 19-k.y. precessional periodicities. Fourth, the picking of cycles is subjective: whether a particular log peak represents a precessionally controlled cycle is not always clear. In an attempt to combat this, cycles were picked from several different logs.
A further difficulty in comparing the cyclostratigraphy to the biostratigraphy is that, in the Miocene, the Berggren et al. (1995b) ages for biostratigraphic dates are not astronomically calibrated. Thus, in Figure 6B, we show the Site 1006 biostratigraphic data with ages that were astronomically calibrated on ODP Leg 154 material from the Ceara Rise (Backman and Raffi, 1997; Chaisson and Pearson, 1997). The ODP Leg 154 tuned ages applied to the Site 1006 bioevents agree well with the predicted ages by the cyclostratigraphy, particularly in the interval 10-12 Ma. However, some astronomically calibrated bioevents (Figure 6B) ~13 Ma deviate more from the predicted sedimentation rate line than the ages of the bioevents as defined in Berggren et al. (1995b). There could be several reasons for this. Use of different taxonomic concepts by individual workers is always a potential for pitfalls. Also, diachroneity of planktonic foraminiferal bioevents is a factor that is not well understood. Therefore, it is impossible to know what has caused the differences and, as a consequence, it is not possible to calibrate the bioevents of Site 1006 because we have used a 'floating' cyclostratigraphy rather than a tuned stratigraphy. The absence of a strong paleomagnetic signal is hampering progress toward a tuned Milankovitch stratigraphy. At present, the cyclostratigraphic record is anchored on one bioevent. Progress could be made by using quantitative micropaleontology in the future to find a range of reliable bioevents to evaluate the position of the cyclostratigraphy with respect to the solar insolation curve (Backmann and Raffi, 1998). Nonetheless, the present cyclostratigraphy is extremely useful for evaluating Milankovitch-induced variations in sedimentation rate.