The following section describes the preliminary shipboard analysis of color alternations (cyclicity) that are observed in cores from Site 1165. This work will be the basis for postcruise studies to more thoroughly explain the observations and processes involved.
As part of standard laboratory operations, cores from Site 1165 were analyzed with the Minolta CM-2002 spectrophotometer for color variations (see "Lithostratigraphy," Fig. F6) and were photographed in color and black and white (see the "Core Descriptions" contents list). The cores were also analyzed for density and magnetic susceptibility variations using the multisensor track (MST) system.
Upon visual inspection, systematic color variations, or color cyclicity, is apparent in the cores from Site 1165. The visual variations are also observed as cyclicity in the lightness factor values from the spectrophotometer (Fig. F54). The L* values show that there is a general downhole darkening of the core that is evidenced by the decreasing trend of the lightness values (see "Lithostratigraphy," Fig. F6). Superimposed on this trend are variations with many wavelengths ranging from centimeter to decimeter size.
In general, the short-wavelength (<3-4 m) lightness variations are caused by cycles between a green to greenish gray facies (lighter, with larger L* values) and a gray to dark gray facies (darker, with smaller L* values; Fig. F55). The greenish facies are structureless diatom-bearing clay with higher biogenic content, dispersed clasts, and lonestones than the dark gray facies, which are mostly clay with some silt laminations (see "Lithostratigraphy"). Greenish facies generally have lower bulk-density, magnetic susceptibility, and organic carbon values.
The lightness cycles occur throughout the hole but are more evident in the upper ~400 mbsf (i.e., green/gray cycles are more evident) than in the lower part of the hole, where rocks are darker and harder with local diagenetic cementation (e.g., opaline silica and carbonate). The ratio of the thickness of the dark to light facies increases with depth downhole (as sedimentation rates also increase; see "Biostratigraphy and Sedimentation Rates"). The cycles can also be seen in the color and black and white core photographs, mostly above ~400 mbsf.
Studies on cores with color variations have been done previously for carbonate-rich sections (e.g., Nobes et al., 1991; Harris et al., 1997; Balsam et al., 1997; Grützner et al., 1997; Kegwin, Rio, Acton, et al., 1998; Ortiz et al., 1999) and have shown that climatic cyclicities (i.e., Milankovitch periodicities) appear in the carbonate and terrigeneous input. In general, climatic cycles are reflected by changes in biogenic productivity and dissolution and by variations in terrigenous sediment fluxes. In low-latitude studies, variations in terrigenous input are caused by changes in sea level and erosion of the continental shelves at times of low sea level. Around the Antarctic margin, carbonate nannofossils and foraminfers are generally uncommon, and the biogenic component consists largely of silicious microfossils. Terrigeneous input variations come largely from glacier erosion and transport of sediments to the continental shelf edge and uppermost continental slope (and then beyond by deep-ocean processes such as turbidity flows and contour currents) during sea-level lowstands when grounded glaciers advance to the continental shelf edge.
Antarctic margin drilling has shown that Milankovitch cyclicities may exist in sediment cores. Upper Miocene and younger sediment drift deposits off the Antarctic Peninsula were drilled during ODP Leg 178 (Site 1095), and color variations were observed, with darker-colored units containing larger amounts of terrigenous debris and lighter-colored beds having more biogenic material. It was noted that "...pronounced cyclicity of the depositional record is evident in visual core descriptions and is recorded to date in color scanner, magnetic susceptibility, gamma-ray attenuation porosity evaluator (GRAPE), and downhole magnetic (GHMT) logging records, as well as (probably) in clay content. Preliminary shipboard analysis has shown this record to contain Milankovitch orbital frequencies [unspecified]...." (Shipboard Scientific Party, 1999, p. 7). At Site 1095, the color and other variations were related to fluctuations of the West Antarctic Ice Sheet. Milankovitch cyclicity has also been reported from drilling activities in front of the Transantarctic Mountains and East Antarctic Ice Sheet in the western Ross Sea by the Cape Roberts Project, Site CRP-2/2A (CRST, 1999; Claps et al., in press; Woolfe et al., unpubl. data). At Site CRP-2/2A, the cyclicites are seen in upper Oligocene/lower Miocene nearshore glaciomarine strata as variations in mean grain size, sand abundance, neutron porosity/density, GRA bulk density, coarse fraction clast abundance, and magnetic susceptibility.
Site 1165 is on the flank of a drift deposit, analogous to ODP Site 1095, but unlike Site 1095, the sedimentation cyclicity is responding to changes in nearby East Antarctica. Additionally, the cyclic green and gray color bands at Site 1165 are of slightly different color than at Site 1095, where brown and gray intervals are present; however, the slight difference may be due to different sediment provenances and geochemical settings of the two sites, in particular the amounts and types of clay and OC present (see "Organic Carbon and Iron").
In order to investigate cyclic changes in sedimentation within lithostratigraphic Units I and II (i.e., above 305 mbsf) of Site 1165, spectral analyses of spectrophotometer color parameter L* were performed in the depth domain for two separate intervals spanning 83-100 mbsf (Fig. F56A) and 107-123 mbsf (Fig. F57A), respectively. For these intervals, almost continuous color (5-cm sampling interval) and GRA bulk-density records (4-cm sampling interval) of high quality were obtained. The spectra were calculated using the Blackman-Tuckey (BT) spectrum algorithm with 50% lags ("Analyseries" program; Paillard et al., 1996). This algorithm (Blackman and Tuckey, 1958) estimates the autocorrelation function from equally spaced data series, weighted by specifically designed windows to discard possible bias, and computes the Fourier transform to obtain the power spectrum. Coherency between lightness and bulk density was obtained by running the program in the cross-spectrum mode. The same data sets were also analyzed with two additional spectral methods. The first (Scargle, 1982) does not require equally spaced series ("Spectrum" program; Schultz and Stattegger, 1997), and the second is a maximum entropy technique (Burg, 1978; see below). The spectra calculated with the three different methods gave similar results for both of the investigated time series.
For the upper interval, significant spectral density peaks in lightness (only BT results are shown) are found at periods of 3.28, 1.45, 1.08, 0.73, and 0.64 m (Fig. F56B). Based on relatively good magnetostratigraphic and biostratigraphic age control, the average sedimentation rate over the interval is 3.5 cm/k.y. Assuming that the sedimentation rate has not changed dramatically over the investigated interval, the cyclicities in depth can be converted to time and the above periods are equal to 93.7, 41.4, 30.9, 20.9, and 18.3 k.y. The 1.45-m period could be suspect because it is close to the section length of 1.5 m, but visual examination of the core photographs confirms distinct color changes that are spaced at ~1.5 m, indicating that the computed cyclicity of 1.45 m is real. The fact that four spectral peaks (93.7, 41.5, 20.8, and 18.2 k.y.) are close to the Milankovitch cycles of 100 k.y. (eccentricity), 41 k.y. (obliquity), 23, and 19 k.y. (precession) suggests an orbital origin of the cyclicity. Although some of the spectral peaks in the gamma-ray attenuation bulk-density spectrum appear shifted when compared to the lightness spectrum, the two records are highly coherent (>90%) in all three orbital frequency bands.
Between 107 and 123 mbsf, significant spectral density peaks were found at 4.27, 1.55, 0.95, and 0.72 m (Fig. F57B). L* and GRA bulk density show a very similar spectral character in this interval, resulting in high coherencies at the significant variance density maxima. Unfortunately, the sedimentation rate for this interval can only be roughly estimated because of the limited resolution of biostratigraphic and magnetostratigraphic dating techniques. The procedure for the detection of orbital cyclicity used for the upper interval is not applicable in this case, but since the Milankovitch cycles are always characterized by a similar hierarchy of periods (although their periods cannot be assumed to be absolutely constant through time) the numerical ratios of the detected peaks can provide indications for orbital control (e.g., Hinnov and Goldhammer, 1991; Fischer et al., 1991). Normalized to the 19-k.y. precessional cycle, the Milankovitch periods have ratios of 1.00:1.21:2.16:5.26. The periods of the above spectral density peaks found in the lightness record of Site 1165 exhibit very similar ratios of 1.00:1.32:2.15:5.93, suggesting an orbital origin for the observed color and density changes. A similar approach has recently been used to detect orbital signals in sedimentary sequences drilled by the Cape Robert Project (Claps et al., in press). Assuming that the 1.55-, 0.95-, and 0.72-m cycles are of orbital origin and are equivalent to the 41-, 23-, and 19-k.y. Milankovitch cycles, refined sedimentation rates can be calculated to be in the range of 3.78-4.13 cm/k.y. for the depth interval from 107 to 123 mbsf.
The variation in the cyclicity with depth was tested for the interval 78-125 mbsf by performing evolutionary spectral analyses over 10-m intervals every 2 m downcore (Fig. F58). The maximum entropy method (MEM) (Burg, 1978) was used with software available from http://www.atmos.ucla.edu/tcd/ssa/. The MEM was chosen because it produces sharp spectral peaks that are easily identified, although the significance of the amplitude values for the peaks is not known.
In the interval between 78 and 100 mbsf, a shift in the spectral peaks can be seen at ~92 mbsf and may be due to a small change in the sedimentation rate. The interval between 107 and 127 mbsf has a consistent spectral pattern that indicates stable sedimentation rates. The correlation of spectral peaks across the core break between 100 and 107 mbsf is difficult to trace because the peak spacing changes, which suggests that sedimentation rate may vary in this interval; however, spectra from around 120 mbsf are quite similar to those from ~98 mbsf, indicating similar periodicities.
Spectral analysis of the lightness and bulk-density measurements demonstrates that the detected frequencies show good consistency with the frequencies predicted by the Milankovitch theory and are reasonably consistent down the section of core examined. The correlation of lightness, GRA bulk density, and magnetic susceptibility (see "Physical Properties") indicates that the color (lightness) data most likely document orbital-driven changes in the marine depositional environment.
Color (lightness) cyclicity appears throughout the hole, although the cyclicity is highly subdued in the darker-shaded rocks of lithostratigraphic Unit III (i.e., below ~305 mbsf) and in the light brown sediments of lithostratigraphic Unit I (i.e., above 63 mbsf). The cyclicity is most apparent in lithostratigraphic Unit II, where biogenic silica concentration of the green beds are >15%-20% (see "Lithostratigraphy," Fig. F6), OC values in dark-colored beds are generally >0.4 wt% (see "Organic Geochemistry," Fig. F53B), and average sedimentation rates are ~3-5 cm/k.y. (see "Biostratigraphy and Sedimentation Rates," Fig. F27). In this depth range (i.e., 63-305 mbsf), the green and gray sediment color-(lightness) may be related to the oxidation state of iron (Fe2+, green) and OC content (>0.3 wt%, gray) (Potter et al., 1980, p. 55). The same relation may hold for the green to gray color- (lightness) banded interval at 20-30 mbsf, which has relatively high OC (0.8 wt%) content and lies within the otherwise uniform brown sediments of Unit I.
Paleomagnetic analyses show that the magnetic intensity is much lower in the gray-green sediments than in the dark-colored beds and that in some intervals of lithologic Units I and II paleomagnetic inclination covaries with the green to dark alternations, especially from ~90 to 130 mbsf. These observations suggest that the variations in the magnetic properties for some intervals are also related to variations in redox conditions, causing partial dissolution of the detrital magnetic minerals and formation of secondary magnetic phases (see "Paleomagnetism," Fig. F38).
At Site 1165, OC values greater than ~0.3 wt% seem to enhance the color (lightness) contrasts but do not alter the location of boundaries for the color (lightness) cycles. Further assessment of the causes of color (lightness) banding is planned for postcruise studies.
The initial identification and evaluation of the cyclicity in color (lightness) of biogenic-rich (greenish gray) and terrigenous (dark gray) intervals suggests that these depositional facies are influenced by Milankovitch periodicities of 100, 41, 23, and 19 k.y. The analysis was done for only two sections of the hole, for late Miocene and early Pliocene times. Similar cyclicities are observed throughout the hole back to early Miocene time, but uncertainties in sedimentation rates for the lower part of the hole at this time preclude determining the times for the cycles. Site 1165 lies in front of the outlet for the Lambert glacier-Amery Ice Shelf system that now drains 22% of East Antarctica with ice and entrained sediment. The color (lightness) cycles at Site 1165 are likely linked to the depositional processes of the onshore glacier system (sediment source) and ocean current system (sediment distribution), which in turn are strongly controlled by climate variations that are driven in part by long-term orbital variations (Milankovitch cycles). Postcruise sedimentological analyses of the cored intervals will supply further fundamental information on compositional and textural variations associated with the observed cylicities.