Modeling Abrupt Climate Change as the Interaction Between Sea Ice Extent and Mean Ocean Temperature Under Orbital Insolation Forcing - Abrupt Climate Change: Mechanisms, Patterns, and Impacts - page 62

stochastic resonance (SR) [Collins et al., 1996] that could
play a role in amplifying the climate system ' s response to
weak external forcing. This is a different mechanism of SR
than what has been usually invoked to explain the D-O
[Ganopolski and Rahmstorf, 2002; Alley et al., 2001]. The
latter is the ampli

contributes to transform bistable noise into a time series that
mimics the data. As illustrated in Figure 3b, a specific level
of noise causes the similarity between the resulting signal
and the insolation to increase, even for very weak astronom-
ical forcing.
The implication of Figure 3, however, is not that the ice
core records are just the result of astronomically modulated
climate noise. Though the similarity between model and
data in Figure 3a would appear to suggest thus, there are

cation of a weak underlying internal
periodic signal by a speci c noise level, while in the former,

the presence of a particular noise level in the system en-
hances the effect of the external forcing (insolation), which

Figure 5. (a) Singular spectral analysis [Ghil et al., 2002] of the NGRIP record. Summary results for other ice cores are
given in Table 1. The original time series was high-pass filtered (corner at 0.000067 c yr 1 ) to eliminate periods at and

longer than the insolation band. Two periods in the millennial D-O range (4.6 and1.67 ka) display high power against the
red noise null hypothesis at the 95% con dence level. (b) Power spectra (multitaper) of an interval of NGRIP where the

in uence of insolation variability should be minimal. The 90% and 95% signi cance levels are marked.