Plio-Pleistocene orbital periodicities in glacially influenced sediments from western pacific margin of Antarctic Peninsula

Summary. Spectral analyses of petrophysical and sedimentological data sets from Plio-Pleistocene sediments cored on the Western Antarctic continental rise during Ocean Drilling Program (Leg 178) show nonharmonic wavelength peaks, with probabilities >90%. When the wavelength peaks are normalized, they exhibit an extremely high correlation factor with predicted Earth's orbital variations for the same time interval. It is also found that both short (~ 95-125 -ky) and long (~400 -ky) eccentricity periodicities clearly emerge from the signal during the whole Pleistocene, without evident switch to obliquity at mid-Pleistocene (~ 0.9 Ma B.P.), as reported in the literature. This suggests that the lithological parameters, proxy of glacial cycles, are controlled, directly or indirectly, by astronomically-forced processes (Milankovitch cycles). Moreover, the constant presence of all orbital periods in the signal since the Lower Pleistocene and the good correlation for the last 2.6 Ma among distant coring sites, based on systematic sedimentological variations at intervals of about 140 and 370 ky, allow to extend the results at regional scale, confirming the hypothesis of the relative stability of the Antarctic ice sheet, at least since Upper Neogene and suggesting that the ice sheet in the Lower-Upper Pliocene and in the Pleistocene could be sensitive to temperature changes. Finally an age correspondence emerges between the correlated events at 1.07, 2.01 and 2.61 Ma, corresponding to Periods of particularly intense ice rafting in the studied area, and relative sea-level highs, suggesting a superimposed eustatic influence. Abstract text. The dynamics of the Antarctic ice sheet is a significant component of the global climate system, so the understanding of its driving mechanisms, still uncompletely known, is of great importance in the analysis of the Earth's climatic evolution. New insights in such dynamics came from studies on the glacial-interglacial sedimentation of upper rise sedimentary drifts off the West Pacific margin of the Antarctica Peninsula. In this sector, the irregular relief of the upper rise is due to large hemipelagic sedimentary drifts, that are separated by nine channel systems formed by turbidity fluxes. Drift sediments alternate from rapidly deposited, poorly fossiliferous silt and clays during glacial intervals, to slowly deposited biogenic muds (with only a minor terrigenous component) during interglacial times (Rebesco et al., 1998, Barker et al., 1999; Barker and Camerlenghi, 2002). Here we present the results from spectral analysis and correlation of Plio-Pleistocene sedimentological (Coarse Fraction) and petrophysical (GRAPE bulk density, Magnetic susceptibility, and Chromaticity a*) data intervals from Sites 1095, 1096 and 1101 ODP Leg 178, drilled respectively from sedimentary drift 7 and 4 (Barker et al., 1999). The spectral analysis program (Horne and Baliunas, 1986; Iorio et al., 1995; Brescia et al., 1995) originally written to deal with unevenly spaced astronomical data sets, is based on a technique which aims to detect the presence and significance of periodicities in unevenly sampled data series. In fact, as it was found in many stratigraphic data sets, the rebinding of the unevenly sampled data to equally spaced bins, and their calculation by means of a conventional periodogram, often alters the perceived frequency and significance of a periodic signal. Among all parameters in the studied intervals, only those with well defined power spectra with more than two wavelengths were considered to record a true cyclic signal (Fig.1). In order to enable comparison from the computed wavelengths expressed in sedimentary thickness (m) and orbital periods expressed in temporal units (ky), the wavelengths of each power spectrum were normalized to their highest well defined frequency, and each orbital calculation for the last 3 Myr computed by Berger and Loutre (1992) was normalized respect to the others (D'Argenio et al., 1998). The two tables of a-dimensional ratios obtained in this way were then cross-correlated, searching for the highest correlation factor. Linear regression analysis of the relative ratios of all wavelengths and orbital periods give correlation factors that are above 0.98, for all power spectra analysed (Fig.1), hence they can be considered strongly linked. Moreover the above data enable a precise computation of the average sediment accumulation rates in the studied intervals (Iorio et al., 2004), which show also a close similarity with both biostratigraphically and palaeomagnetically computed sedimentation rates (Iwai et al., 2002; Acton et al., 2002). Looking at the highest energy peaks (bold marked in Fig. 1), we find out that the short eccentricity (~95 and 123-ky) strongly influences the Lower-Upper Pliocene (Site 1095) and the whole Pleistocene of Sites 1095, 1101 and 1096. The ~400-ky long eccentricity strongly emerges from the Upper Pliocene-Pleistocene sedimentary record of Sites 1095 and 1101. Moreover the ~20 ky precession is evident in the Upper Pliocene-Pleistocene of Sites 1096 and 1101. Finally the ~64 ky period characterizes the Lower-Upper Pliocene and Mid Pleistocene of Sites 1095, 1096 and 1101. In order to allow a comparison of the FFT analysis results among all three Sites data sets, all sedimentological and petrophysical data, which were collected at different intervals in each site, were smoothed with an average window of 140 cm (which is the longest sampling interval). The smoothed data were calibrated in Ma (Fig.2 ) using the sedimentation rates astronomically computed (Iorio et al., 2004) as well as biostratigraphically and paleomagnetically constrained (Iwai et al., 2002; Acton et al., 2002) in the intervals excluded from FFT analysis. Subsequently, the variations, of all smoothed and calibrated parameters, syncronously (±0.05Ma) occurring in at least two Sites belonging to two distinct sedimentary drifts (drift 7 and 4), were considered to be an expression of the same event, (horizontal lines in Fig 2) and the averaged age was computed. When the same event was recorded in all three Sites, it was considered an even stronger event, occurring at 0.39, 0.9, 1.07, 1.42, 2.01, 2.13, 2.61 Ma) (stars in Fig.2).The average time interval between two subsequent events during the last 2.6 Ma, was found of 0.14±0.04 Ma, and of 0.37±0.17 Ma for the strongest events recorded in all three Sites. So these time-intervals are close to that computed for the Plio-Pleistocene power spectra of higher energy wavelengths (Fig.1). Finally the constant presence of all the orbital periodicities in the signals since Lower Pleistocene and in part of the Lower-Upper Pliocene of Site 1095, confirms the relative stability of the Antarctic ice sheet since Upper Neogene (Barker and Camerlenghi, 2002;) and implies its sensitivity to the insolation changes during the studied time interval.