Geoscience Reference
In-Depth Information
probits of
Epistomina mosquensis
abundance data for Tojeira 1 section.
Right side
: Same with
observations and spline curve for Tojeira 2 section superimposed. Patterns were slid with respect
to one another until a reasonably good fit was achieved. Zero distance (at sample 6.29 in Tojeira 1)
falls just below base of overlying Cabrito Formation (cf. Fig.
9.15
). Correlation between the two
sections is poorest along the 35 m data gap in Tojeira 2 (Source: Agterberg
1990
, Fig. 3.11)
9.3 Large-Scale Applications of Ranking and Scaling
In applications to real datasets, attention has to be paid to the frequency
distributions of single events as well as to the two-dimensional frequency distribu-
tions of pairs of events included in the dataset. This is because of the prevalence
of missing data in practical applications. Biostratigraphic exploratory well data
primarily consist of last occurrences (LOs) of taxa, because the sampling procedure
normally consists of taking cuttings at regular intervals during the drilling. As
mentioned before, lower down a well, material from higher up may contaminate
later cuttings so that observed locations of FOs can become biased and, therefore,
should not be used. The following real-data examples of cumulative event fre-
quency distributions are for Cenozoic microfossils from (A) 30 North Sea wells,
using 1,430 event records of 289 taxa of benthic and planktonic Foraminifera,
miscellaneous shelly microfossils, and dinoflagellates (from Kaminski and
Gradstein
2005
), and (B) 27 Labrador and northern Grand Banks wells, using
960 event records of 178 taxa of benthic and planktonic Foraminifera, and miscel-
laneous other microfossils (from Gradstein and Agterberg
1982
). Figure
9.17
shows
cumulative frequency distributions for all events in datasets (A) and (B). These
Search WWH ::
Custom Search