A detailed study of the periodicity of the color reflectance requires an accurate age scale. Clearly, to avoid circular reasoning, the reflectance record itself must not be used to derive this scale by orbital tuning. We therefore turn to magnetic susceptibility to explore the potential for tuning and matching to the global oxygen-isotope record. Magnetic susceptibility was measured on whole cores from the three holes as part of the multisensor track analysis that is standard on board the JOIDES Resolution. The susceptibility was quite low, usually between 1 x 10–5 and 6 x 10–5 (volume susceptibility in SI units; (Fig. 4). Spuriously low values resulted from the numerous gas expansion voids present within the cores. Such expansion tends to be more common toward the bottom and top of a core. Thus, there may be a tendency for artificial cycles, generated by core breaks, within any one hole.
The available values were first reduced to a sampling interval of 8 cm (resolution near 1 k.y.), which was considered sufficient for the purpose of testing for cycles with periods >10 k.y. Values that had a range more than twice that of their neighbors in the previous 40 cm, counting downward in a core, were eliminated or reduced unless visual inspection showed similarly extreme values within the next 40 cm downcore. After thus despiking each record, it was smoothed by a five-point Gaussian filter. The in-core depth of each averaged value was then adjusted by using the identical filter.
As previously mentioned, in terms of providing for accurate measurements, the most valuable portion of each core is the mid-section. To produce a stacked record, we first aligned the characteristic features of individual core records from different holes (see "Composite Section" section, "Explanatory Notes" chapter, this volume; Hagelberg et al., 1992). The standard ODP depth scale can thus be converted into a meters composite depth scale, which represents the common depth scale for all three holes at one site. Coring gaps and sediment expansion during core retrieval result in a composite depth that is expanded compared with the driller's depth by about 10%. Each core was assigned a sinusoidal weighting function set to 0.03 at the top and bottom and to unity at the center. All individual magnetic susceptibility values, together with their weights, were then sorted for all three holes according to their assigned composite depths. Adjacent values in the sorted series were averaged three at a time, weighting each average according to the weights of the individual values for magnetic susceptibility being summed. The composite depth values were treated in an identical fashion. The end result of this procedure is a series of magnetic susceptibility values ordered according to composite depth and with information strongly biased toward measurements at the center of stacked cores.
Earlier work in the region around Site 1075 and preliminary estimates based on core-catcher examination suggested a sedimentation rate near 200 m/m.y. in the uppermost portion of the site, and one of about 100 m/m.y. overall. After an initial rough match of selected peaks and valleys of the magnetic susceptibility record to an oxygen-isotope record (described below), an equation describing total sedimentation rate as a function of composite depth was applied to the stacked record as follows:
Based on this equation, preliminary ages were assigned to all composite depth values. The ages were then adjusted (mostly by increasing them between 12% and 20%) by matching the magnetic susceptibility curve to the oxygen-isotope record of Ontong Java Plateau (Berger et al., 1994; revised for the last 800 k.y. in Berger et al., 1996). This record is based on analysis of Globigerinoides sacculifer and is thought to represent sea-level change (having very little precessional component, in the middle of the tropics). The procedure yielded a depth for the Stage-20-to-Stage-19 transition at 82 meters below seafloor (mbsf), which agrees with the depth assignment for the Brunhes/Matuyama boundary (see "Paleomagnetism" section, "Site 1075" chapter, this volume). Considerable adjustment of initial age assignments was necessary for the uppermost cores from Site 1075, which appear to be greatly expanded. Amplitudes of magnetic susceptibility seem unusually high and were reduced by a factor of 2 for this section (~0–50 ka).
The match of magnetic susceptibility data to the oxygen-isotope curve (OJsox96; Ontong Java G. sacculifer oxygen isotopes, 1996) is excellent between 1.1 and 0 Ma (Fig. 5). This is especially true if the susceptibility series is partially integrated (labeled "partially integrated susceptibility" in Fig. 5A). The integration is by exponential decay of each value, from older to younger ages, with an e-folding time of 7000 yr. If this curve is the one to be correlated with oxygen isotopes, the entire scale needs to be adjusted by adding ~6000 yr. We have not done this, but this option should be kept in mind when considering phase relationships of the various cycles.
For the sediments older than about 1 Ma, it proved extremely difficult to find a satisfactory match, even after considerable experimentation. In the following text we will, therefore, largely restrict discussions to the periods within the last 1.2 m.y.; that is, the Milankovitch Chron (625–0 ka) and the Croll Chron (1240–625 ka; time scale of Berger et al., 1994).
The excellent fit of the magnetic susceptibility data to the oxygen-isotope curve should result in the appearance of Milankovitch power in the cycles of susceptibility. Fourier analysis of the autocorrelation series of OJsox96 (see Fig. 6) extracts the expected periodicities for the Milankovitch and Croll Chrons. The Milankovitch Chron shows the strong dominance of the ~100-k.y. cycle (with a peak near 98 k.y. related to eccentricity and internal oscillation) and a modest but clear 41-k.y. cycle (related to the changing obliquity of the Earth's axis). The Croll Chron does not show the 100-k.y. cycle but shows power near 80 and 140 k.y.
The cycles in the magnetic susceptibility record resemble those in the oxygen-isotope record, but they are by no means identical. The analysis yields strong periodicity near 100 and 41 k.y. for the last 500 k.y. (Fig. 7, bottom curve). The power near 70 k.y. represents the beat between these two periods (beat frequency = difference between the two base frequencies; that is, 1/41 – 1/98). The next older section (500 ka–1 Ma) shows only the 41-k.y. cycle (weakly) and a strong multiple of this cycle near 83 k.y. The spectrum of the next older section (1.0–1.5 Ma) suggests that the instantaneous sedimentation rate is overestimated here by about 10% (from the offset of the peak expected at 41 k.y.). The biostratigraphic data confirm this (see "Biostratigraphy and Sedimentation Rates" section, "Site 1075" chapter, this volume). The oldest section (1.5–2.0 Ma) seems strongly dominated by the 41-k.y. cycle.