CONCLUSIONS

We used wavelets to analyze the frequency content of core and log measurements from upper Miocene to Pleistocene drift sediments located west of the Antarctic Peninsula. To assess the reliability of our results and their relation to insolation, we also analyzed insolation and truncated insolation signals transformed to the depth scales of Holes 1095B and 1096C. We found that accurate dating of the cores and logs is very important for accurate determination of the dominant frequencies, and hence for the climatic information implied by those frequencies. Our results show that wavelet transform analysis can be useful for the stratigraphic interpretation of core and log measurements.

Some main results that relate to the understanding of the glacial history of the Antarctic Peninsula can be highlighted:

1. A few of the log and core measurements vary with periods close the Milankovitch periods, not only in the Pliocene-Pleistocene, but also in the Miocene. For example, a persistent ~100-k.y. cycle (with variable amplitude) is present in some the wavelet transforms for the Pliocene-Pleistocene period.
2. Periods different from those of Milankovitch are revealed, such as 30-35, 50-55, and 74-78 k.y. Similar periods have been observed by other authors for the same time interval (Imbrie and Imbrie, 1979; Imbrie, 1985; Ghil, 1987; Robinson, 1990; Rea, 1994; Bolton et al., 1995; Lauer-Leredde et al., 1998).
3. Discontinuities in the frequency content of the log and core measurements are recognized in the depth-wavenumber maps. A change is observed during the late Miocene, at ~8 Ma, giving rise to dominantly shorter periods until 7.5 Ma. It is important to note that this pattern correlates with higher sedimentation rates, whatever the considered age models. The Pleistocene-Pliocene interval also exhibits two dominant shifts: the first near 2.1-2.8 Ma and the second near 3.5-3.8 Ma. This is consistent with changes in lithology and climatic transitions elsewhere on the globe (Liu and Chao, 1998).