Teresa King Hagelberg

Milankovitch band forcing of sub-Milankovitch climate variability during the Pleistocene

Climate variability at periods from 10 kyr to 12 kyr that originates from Milankovitch band forcing is quantified at three locations for the late Pleistocene (eastern equatorial Pacific Ocean Drilling Program site 846, eastern equatorial Atlantic ODP site 663, and northeastern Atlantic Deep Sea Drilling Project site 609). Variability at these periods is not present in the primary Milankovitch forcing, so no linear linkage to Milankovitch band variations is possible. However, these periods are equal to harmonics of precession band oscillations. The magnitude of interactions between processes that occur at different timescales can only be resolved in time series data through application of higher-order statistics. Through such an application, we demonstrate that up to 75% of the variance in the 10- to 12-kyr band in the sediment records is nonlinearly transferred from precession band (19–23 kyr) variations. Within the millennial to sub-Milankovitch band, defined as the band of variance ranging from ∼15 to ∼2 kyr, approximately 1/3 of the variability in the records studied is consistent with a low-frequency, Milankovitch band origin. This variability may derive from high sensitivity of the tropics to summertime insolation in both hemispheres relative to wintertime insolation. A mechanism having equatorial origin and related to low-latitude precession variations appears consistent with the observations. Because the phase coupling between 10- to 12-kyr oscillations and precession is resolved, this result has implications for development of models which seek to explain global climate variations on this timescale.

Linear and nonlinear couplings between orbital forcing and the marine δ18O record during the Late Neocene

Previous investigations of the response of Plio-Pleistocene climatic records to long-term, orbitally induced changes in radiation have considered a linear response of climate. While the second-order statistics of power spectra and cross spectra provide necessary information on linear processes, insight into the nonlinear characteristics of Pliocene and Pleistocene climate is not provided by these statistical quantities. Second-order statistics do not contain the phase information necessary to investigate nonlinear, phase-coupled processes. Such information is provided by higher-order statistical quantities. In particular, bispectral analysis indicates that nonlinear couplings are present in the climatic (radiative) forcing at the Milankovitch frequencies. Through a linear transfer, this forcing produces similar nonlinear couplings in deep-sea sedimentary oxygen isotope records (ODP site 677 and DSDP site 607) from 1.0 to 0 Ma during the late Neogene. This analysis suggests that during the late Pleistocene, the dominance of the 100,000 year cycle in the climate record is consistent with a linear, resonant response to eccentricity forcing. In the period from 2.6 to 1.0 Ma, a change in the nature of the climatic response to orbital forcing is indicated, as phase couplings present in the isotopic time series are not similar to the phase couplings present in the insolation forcing. Third-order moments (skewness and asymmetry) are used to quantify the shape of the climatic response. From 2.6 Ma to present, an increase in the asymmetry (sawtoothness) of the oxygen isotopic records is accompanied by a corresponding decrease in the skewness (peakedness) of the records. This indicates an evolution in the nature of the phase coupling within the climate system. These results may provide important constraints useful in development of models of paleoclimate.