Geoscience Reference
In-Depth Information
to changes in seawater level also observed in Portugal now at the other side of the
Atlantic Ocean, which originated during the Jurassic.
In order to use RASC ranking or scaling results for correlation between strati-
graphic sections in CASC, it is necessary to project their positions in the regional
standard, which is an average based on sections, back onto the lines for individual
lines. This topic already has been discussed before in this chapter. Figure 9.25
illustrates how the indirect method previously explained in Sect. 9.2.2 (Fig. 9.14 )
was used in earlier versions of CASC. In post-1998 RASC-CASC computer pro-
grams (e.g., the current Version 20) several simplifications were made including
that the smoothing factor for the step of event levels to depths (Fig. 9.25b ) was set to
zero so that the level-depth curve passes through the data points. Because of
differences in rates of sedimentation through time, error bars (e.g., 95 % confidence
intervals), which are assumed to be symmetrical along the scales constructed for
ranking or scaling, can become asymmetrical after projection onto individual wells.
The degree of asymmetry depends of the curvature of the final curve for probable
positions of the events in each individual well. Anther post-1998 simplification is
that lines of correlation in scattergrams, in which observed events in wells are
plotted against optimum sequence, are fitted as downward decreasing quadratic
curves by least squares. The main reason for these simplifications was that the
original method (Fig. 9.25 ) had to be applied separately to all individual wells. This
procedure turned out to be very time-consuming in practice. The use of quadratic
curves as shown previously in Fig. 9.6 turned out to be a good and fast substitute.
Figure 9.26 is an early example of a CASC multiwell comparison produced by
means of the original CASC program (Agterberg et al. 1985 ). The underlying scaled
optimum sequence was based on 54 last occurrences of Cenozoic Foraminifera in 7 or
more wells (out of a set of 21 NWAtlanticMargin wells). The CASC version used had
an additional step in that cumulative RASC distances were transformed into millions
of years on the basis of a sub-group of 23 Cenozoic foraminiferal events for which
literature-based ages were available. Because of significant uncertainties associated
with this extra step, long-distance correlations in later versions of CASC are based on
the ranked or scaled optimum sequence only. Three types of error bar are shown in
Fig. 9.26 . A local error bar is estimated separately for each individual well. It is two
standard deviations wide and has the probable isochron at its center. Use is made of the
assumption that rate of sedimentation is linear in the vicinity of each isochron
computed. Consideration of variable sedimentation rates results in the asymmetrical
modified local error bar of Fig. 9.26b . Like a local error bar, the global error bars of
Fig. 9.26c are symmetric but they incorporate uncertainty in age derived from
uncertainties in RASC distances for all 54 foraminiferal events in the scaled optimum
sequence based on all (21) wells.
Figure 9.27 shows correlation of ten Cenozoic isochrons between six wells on the
Grand Banks and Labrador Shelf including the three wells used for Fig. 9.26 .Inthis
study performed by Gradstein and Agterberg ( 1985 ) CASC-derived positions are
compared with observed depths. Conventional chronostratigraphic correlation only
uses observed depths. The CASC-based depths result in slight up or down adjustments
of the age boundaries. The data used to obtain Figs. 9.26 and 9.27 was published as the
Gradstein-Thomas database in the Appendix of Gradstein et al. ( 1985 ).
Search WWH ::




Custom Search