|
|
|
|
|
|
|
|
|
|
|
|
|
 |
|
|
|
|
Milankovitch orbital cycles (eccentricity at periods of 95K, 123K, and 410K; obliquity at 41K; and precession at 19K and 23K) are expected in logs of pelagic sediments for the following reasons:
- Orbital changes cause global or regional climatic changes.
- Climatic changes affect the mineralogy or porosity of the sediments.
- Logs detect mineralogy and porosity changes.
The cycles may appear on different logs in different regions (so far we have had success with gamma, density, resistivity, magnetic susceptibility, and sonic logs), and it is possible that different logs from the same well will show energies at different Milankovitch periods.
The climate system also varies on sub-orbital time scales, and this climate variability is similarly reflected in the composition and physical properties of the sediments. In regions where the sedimentation rate is high enough, or conversely in logs with sufficient vertical resolution, the millennial scale variability can also be documented.
There are several prerequisites to successfully identifying any climate cycle through spectral analysis. First, log display makes a difference. A log plot that shows broad compaction trends may obscure fine-scale Milankovitch cycles. Second, accurate sedimentation rates are needed for confirmation that any detected periodicity is at Milankovitch frequencies. Reversal stratigraphy gives more accurate sedimentation rates than paleontology, because the latter has errors at both datums that blow up when calculating a sedimentation rate. Lacking precise sedimentation rates, one will need to detect at least two Milankovitch periods (preferably in more than one log) before any confidence can begin to be placed in them. Third, a high sedimentation rate is needed for logging tools with 0.5 vertical resolution to detect high frequencies (e.g. 19K, 23K, and 41K). Fourth, beware of cycles caused by local sedimentary phenomena (e.g. turbidites) rather than climate: the depth period of the latter will change with the sedimentation rate, but not the former. Keep in mind that the 41K cycle is the only truly constant period for all ages. Eccentricity strength varies somewhat between 95K and 123K as a function of time, and precession strength varies between 19K and 23K (although ideally one would find separate eccentricity peaks and separate precession peaks). Also, the shortest periods are the most likely to be smeared by small changes in sedimentation rate within a log interval.
Spectral analysis is the most common means of characterizing periodicity in logs and can be undertaken with either depth or age as the independent variable. Ultimately, however, a conversion from depth scale to age must be performed in order to understand the driving forces behind the variability. There are a large number of programs available on various platforms that easily allow spectral analysis to be performed. Perhaps the easiest and most commonly employed (and accepted) method is to use the Macintosh program Analyseries to perform the analysis. In order to generate power spectra in Analyseries you need to do the following:
- Generate a tab delimited text data file for the mac with the first column as depth (or age)
- Import it to Analyseries (open it from within the application)
- Select (click on) the data you wish to analyze and choose a method from the "Math" menu. Blackman Tukey is the most common method used by geologists/paleoceanographers, but a variety of methods should be compared to insure that the results are robust. The Blackman-Tukey method is nice because, unlike some other methods, it gives confidence estimates for the results.
- The resulting (frequency vs. power spectra) output can be copied and pasted into any spreadsheet program or plotted directly in Analyseries.
|
Example of spectral analysis figure. (Click to enlarge.)
|
|
Once Milankovitch cycles have been positively identified in the logs, the regular pacing of these records can be used to refine or "tune" the timescale by correlating the climate driven cycles to the astronomical forcing. Such correlations are capable of producing much more accurate and highly resolved age models than are obtainable by other methods. Furthermore, with a high-resolution timescale, it becomes possible to make phase estimates for the relative responses of the different components of the climate system, and to determine the rates of various geologic processes.
The SAGAN program, designed to correlate core and downhole log records, is also capable of automatically (by maximizing coherence) or manually (by graphical selection of tie points) correlating logs to insolation records. In this way it is possible to generate a highly resolved age-depth model for a well in a matter of minutes. However, any age model generated by tuning in this fashion should be considered tentative because although a high coherence is good for estimating the success of tuning, it is not necessarily an indication of the degree of common amplitude modulation (a basic test of whether the tuning is correct). Instead, other methods such as complex demodulation, which assesses the relationship between amplitude modulation in both the data and the inferred forcing, are necessary to evaluate the validity of the timescale [Shackleton et al., 1995].
|
|
|
