Introduction - Rock Magnetic Cyclostratigraphy - page 4

1.2 Basic Steps of a Rock Magnetic
Cyclostratigraphy Study
The steps to a rock magnetic cyclostratigraphy are summarized in Figure 1.1.
The first step in conducting a rock magnetic cyclostratigraphy is to select the
stratigraphic section for study and estimate its sediment accumulation rate,
so that the correct sampling interval can be chosen. The frequencies that can
be detected by the time series analysis are limited by the Nyquist frequency .
The shortest cycle that can be observed by time series analysis must be
sampled at least twice per cycle; therefore, if precession (nominally a 20 kyr
period) is to be captured, the rocks should be sampled at least once every
10 kyr of stratigraphic thickness. In some cases, previous work, either biostra-
tigraphy, magnetostratigraphy , or geochronology of ash layers, can be used
to calculate the sediment accumulation rate. In most cases, though, only an
estimate of the sediment accumulation rate, based on the rock's lithology and
depositional environment, is available. Sadler's (1981) comprehensive study
of sedimentary record completeness can be an important source of these
estimates. However, sampling precession twice per cycle is only a bare
minimum. Aliasing could occur if shorter cycles present in the data are
undersampled, producing apparent longer period cycles which are only an
artifact of the sampling interval. It may be better to target possible precession
cycles with three or four samples per cycle.
Once the sampling interval is selected, unoriented rock samples, ~4-5 cm
in size, are collected throughout the section. If the modulation of precession
(~20 kyr) by short (~100 kyr) and long (405 kyr) eccentricity is the desired
target for detection, then the sampling interval is at least a 10 kyr stratigraphic
thickness of sediment (ideally 5-7 kyr), and the stratigraphic thickness of
the sampled interval should be long enough to record about six repetitions
of the longest period cycle (Weedon 2003). Furthermore, the longer the
series, the better the bandwidth resolution, i.e., the narrower the spectral
peaks and hence the better one can resolve the frequency of individual
cycles. If long eccentricity is the longest period targeted for detection, then
the section should be at least ~2-2.5 million years long. For typical hemipe-
lagic marine sediments with sediment accumulation rates of about 10 cm/
kyr (Sadler 1981), these requirements would mean a sampling interval of
about 0.5-1 m and a section thickness of at least 200-250 m, generating a
minimum of 200-250 samples. For best results, at least three or four samples
should be collected for every precessional cycle, increasing the number
of samples to 500 samples for the 2- to 2.5-million-years long section.
The work needed to collect this large number of samples is offset some-
what because the samples do not need to be oriented, as would standard
paleomagnetic samples, because only the intensity of the sample is mea-
sured, not the direction of its magnetization. Sampling a 250 m section at
75 cm intervals would take about 3-4 days in the field.