Geology Reference
In-Depth Information
step (AR(1) process: first-order autoregressive process), i.e., it is a way of
modeling the natural inertia, or hysteresis, in a system and is reasonable for
geological and climate processes. The MATLAB script
rednoise.m
in the
Appendix can also be used to generate an AR(1) red noise spectrum for
identification of significant spectral peaks in the MTM power spectrum.
7.5.4
Evolutionary Spectrogram
All stratigraphic data series suffer from uneven sediment accumulation rates
and hiatuses in the record. Furthermore, rock magnetic cyclostratigraphic
series may suffer from uneven quality of the recording of natural periodic
behavior, i.e., global climate cycles, due, for example, to postdepositional
changes in the magnetic mineralogy or variations in the supply or transport
of magnetic minerals to the depositional basin. One good way to investigate
whether these processes have affected the rock magnetic cyclostratigraphic
data series is to calculate an evolutionary spectrogram (Figure 4.24) using
MATLAB script
evofft.m
(Appendix). (Analyseries will also calculate evolu-
tionary spectrograms.) The script calculates the periodogram of a portion
(window) of the data series and steps the window through the series, plotting
the resulting spectra as a function of the step increment. The user selects the
data window size, which can be no larger than the total data series but should
be much smaller, and the increments at which the window is stepped through
the data series that are much smaller than the data window. The rock
magnetic data evolutionary spectrogram can be compared to an evolutionary
spectrogram generated from the theoretical insolation series of similar age
(Laskar et al. 2004) to see if changes in the amplitudes of Milankovitch
spectral peaks match (Figure 4.24) thus providing evidence to support
identification of astronomically forced climate cycles.
7.5.5
Tuning and Filtering
As indicated in Chapter 4, filtering can be an important tool for examining
the frequency content of the rock magnetic cyclostratigraphic data series.
For instance, filtering the data series at a suspected precessional frequency
and checking its amplitude modulation for eccentricity frequencies is a use-
ful way to identify Milankovitch cycles in the data (see the Cupido Fm. and
Daye Fm. studies in Chapter 6). Furthermore, filtering a data set at a given
Milankovitch frequency can facilitate tuning to the theoretical insolation
variations for that Milankovitch frequency at the age of the sedimentary
sequence. The tuning can help remove the effects of slight changes in sedi-
ment accumulation rate and small hiatuses in the record, with the potential
of enhancing the power spectrum. Analyseries can apply Gaussian band
pass filters to a data series;
gaussfilter.m
and
tanerfilter.m
(Appendix) provide
stable alternatives.
