628
von Dobeneck and Schmieder
extraordinarily high bispectral intensities of the intermodulation frequencies (1/41 + 1/23-1119) kyrl
and (1141-1123+1/19) kyrl. Both combination
tones express nonlinear interaction of obliquity
with the envelope ofthe composed precession signal, which has a frequency of (-1123+1119) kyrl,
corresponding to an (averaged) eccentricity frequency of 11109 kyrl. This interaction provides the
trigger mechanism for the late Quaternary 100 kyr
ice age cycles.
Millennial climate cycles are believed to result
from free, self-sustained glaciomarine oscillations
as proposed by many coupled climate models (e.g.,
Birchfield et al. 1994; Paillard 1995). Large millennial temperature oscillations documented by ice
core 8 18 0 studies from both hemispheres (Dansgaard et al. 1993; Benderet al. 1994) have recently
also been found in various marine high-resolution
proxy records (Behl and Kennett 1996; Charles et
al. 1996; Moros et al. 1997; Adkins et al. 1997),
with some reservations also at the Ceara Rise
(Curry and Oppo 1997). It is therefore not overly
ambitious to tentatively correlate the millennial
signalcomponentofGeoB 1523-1 (Fig. 16d) with
the GRIP 8 18 0 record (Fig. 17). Although both
patterns are not exactly mirror images, the similarity is quite reasonable given the fact that a 2.8 m
and a 2800 m record from two climate extremes
are compared.
Conclusions
The SUSAS and CEARIS studies demonstrate
how the methods of cyclostratigraphy and environmental magnetism can be fruitfully combined.
Other than most physical properties parameters,
magnetic susceptibility depends linearly on the
concentrations of a few (iron-bearing) minerals,
predominately (titano-) magnetite. In most marine
environments its mean value and climatic modulation primarily mirrors the ratio of terrigenous and
biogenic sediment accumulation. Susceptibility
variations can be closely related to carbonate dissolution cycles as in case of the SUSAS cores, but
may also primarily respond to climatic cycles of
terrigenous supply as in the CEARlS example. The
basic mechanisms of these two cycles are well
studied and largely understood. They are effects
of deep water formation, sea level change, aridity
and wind intensity, all intimately linked to orbital
cycles. Although the phase relations of8 18 0 and K
differ regionally, the orbital response of susceptibility to climatic change is often adequate for
cyclostratigraphy and orbital tuning. Pilot spectral
studies on selected cores yield applicable phase lags
and indicate best tuning targets.
Particularly in low-accumulation environments
phase settings are not as important as pattern reproducibility. Because susceptibility logging generates a complete low-noise image based on the
sediment volume, subtle signal characteristics are
resolved enabling pattem identification even in problematic sections. Aperiodic trends, shifts and amplitude fluctuations of the susceptibility record
document modifications of the depositional system.
Extracted by lowpass filtering and evolutionary
spectral analysis, these features may have more
interesting climatic implications than a perfect orbital correlation scheme. In the SUSAS example,
analogue long term trends of subtropical South
Atlantic K and %CaC0 3 and global 8 13 C (but not
(180) records indicate marked changes at around
0.95 and 0.6 Ma. Repeated converse shifts of
100 kyr and 41 kyr spectral power relative to an
ETP model substantiate pC0 2 threshold concepts
(e.g., Saltzman and Verbitsky 1993; Raymo 1997)
of the mid-Pleistocene transition.
In addition, most high-resolution susceptibility
records carry important signal components in the
sub-Milankovitch and millennial bands. In case of
the eight CEARlS records these high-frequency
variations are largely coherent and can be used to
improve the internal precision of the combined
cyclostratigraphic age model. As shown by spectral and bispectral analysis, the signal modulation
in the 7-18 kyrband is mainly due to periodic harmonics and combination tones of orbital frequencies such as the doubled precession cycle predicted
for tropical climate by models (Short et al. 1991).
Similar spectral interpretations in the < 7 kyr band
are inhibited by the complexity of combination tones
and uncertainty of the age model. A tentative correlation of millennial CEARIS variations to
Dansgaard-Oeschger cycles (Dansgaard et al.
1993) yields quite convincing results within the limitations of both age models. Both categories of high-
von Dobeneck and Schmieder
extraordinarily high bispectral intensities of the intermodulation frequencies (1/41 + 1/23-1119) kyrl
and (1141-1123+1/19) kyrl. Both combination
tones express nonlinear interaction of obliquity
with the envelope ofthe composed precession signal, which has a frequency of (-1123+1119) kyrl,
corresponding to an (averaged) eccentricity frequency of 11109 kyrl. This interaction provides the
trigger mechanism for the late Quaternary 100 kyr
ice age cycles.
Millennial climate cycles are believed to result
from free, self-sustained glaciomarine oscillations
as proposed by many coupled climate models (e.g.,
Birchfield et al. 1994; Paillard 1995). Large millennial temperature oscillations documented by ice
core 8 18 0 studies from both hemispheres (Dansgaard et al. 1993; Benderet al. 1994) have recently
also been found in various marine high-resolution
proxy records (Behl and Kennett 1996; Charles et
al. 1996; Moros et al. 1997; Adkins et al. 1997),
with some reservations also at the Ceara Rise
(Curry and Oppo 1997). It is therefore not overly
ambitious to tentatively correlate the millennial
signalcomponentofGeoB 1523-1 (Fig. 16d) with
the GRIP 8 18 0 record (Fig. 17). Although both
patterns are not exactly mirror images, the similarity is quite reasonable given the fact that a 2.8 m
and a 2800 m record from two climate extremes
are compared.
Conclusions
The SUSAS and CEARIS studies demonstrate
how the methods of cyclostratigraphy and environmental magnetism can be fruitfully combined.
Other than most physical properties parameters,
magnetic susceptibility depends linearly on the
concentrations of a few (iron-bearing) minerals,
predominately (titano-) magnetite. In most marine
environments its mean value and climatic modulation primarily mirrors the ratio of terrigenous and
biogenic sediment accumulation. Susceptibility
variations can be closely related to carbonate dissolution cycles as in case of the SUSAS cores, but
may also primarily respond to climatic cycles of
terrigenous supply as in the CEARlS example. The
basic mechanisms of these two cycles are well
studied and largely understood. They are effects
of deep water formation, sea level change, aridity
and wind intensity, all intimately linked to orbital
cycles. Although the phase relations of8 18 0 and K
differ regionally, the orbital response of susceptibility to climatic change is often adequate for
cyclostratigraphy and orbital tuning. Pilot spectral
studies on selected cores yield applicable phase lags
and indicate best tuning targets.
Particularly in low-accumulation environments
phase settings are not as important as pattern reproducibility. Because susceptibility logging generates a complete low-noise image based on the
sediment volume, subtle signal characteristics are
resolved enabling pattem identification even in problematic sections. Aperiodic trends, shifts and amplitude fluctuations of the susceptibility record
document modifications of the depositional system.
Extracted by lowpass filtering and evolutionary
spectral analysis, these features may have more
interesting climatic implications than a perfect orbital correlation scheme. In the SUSAS example,
analogue long term trends of subtropical South
Atlantic K and %CaC0 3 and global 8 13 C (but not
(180) records indicate marked changes at around
0.95 and 0.6 Ma. Repeated converse shifts of
100 kyr and 41 kyr spectral power relative to an
ETP model substantiate pC0 2 threshold concepts
(e.g., Saltzman and Verbitsky 1993; Raymo 1997)
of the mid-Pleistocene transition.
In addition, most high-resolution susceptibility
records carry important signal components in the
sub-Milankovitch and millennial bands. In case of
the eight CEARlS records these high-frequency
variations are largely coherent and can be used to
improve the internal precision of the combined
cyclostratigraphic age model. As shown by spectral and bispectral analysis, the signal modulation
in the 7-18 kyrband is mainly due to periodic harmonics and combination tones of orbital frequencies such as the doubled precession cycle predicted
for tropical climate by models (Short et al. 1991).
Similar spectral interpretations in the < 7 kyr band
are inhibited by the complexity of combination tones
and uncertainty of the age model. A tentative correlation of millennial CEARIS variations to
Dansgaard-Oeschger cycles (Dansgaard et al.
1993) yields quite convincing results within the limitations of both age models. Both categories of high-