Geoscience Reference
In-Depth Information
Spatial variation in cross-set thicknesses
easily revealed by an exceedence frequency plot
(Drummond & Wilkinson, 1996). The exceedence
frequency,
EF
(
x
i
), of a particular value
x
i
of the
measured variable (
x
) is defined as the number of
data (
n
i
) with values greater than
x
i
, divided by the
total number of data (
n
):
The Student
t
-test and the non-parametric RAM
and RUD tests and Spearman rank-correlation test
(Davis, 2002; Longhitano & Nemec, 2005) failed to
reveal any significant spatial trends in cross-set
thicknesses. The differences in the cross-set thick-
ness variances for the sandstone ridges, both within
and between the individual wells, also appear to be
insignificant at ≥ 90% confidence level. This means
a lack of significant differences in cross-set thick-
nesses between the sandstone ridges and also
between their central and lateral parts. The thickest
sections of sandstone ridges encountered by the
wells show some irregular upward thinning of
dune cross-sets, which probably reflects an increase
in the erosional impact of waves in an aggrading
tidal sand ridge. There is little textural difference
between the thicker and thinner cross-sets, which
supports the notion of erosional truncation.
The tidal sand ridges were laterally extensive
and had a gentle relief, as the inter-ridge sediment
accumulation kept pace with the ridge growth
(Fig. 10). The reported mean side slopes of mod-
ern sand ridges are mainly < 1º (McBride, 2003).
The lateral bathymetric differences were reduced
at aggradation maxima, with the impact of waves
increasing and becoming more uniform. This
subtidal sea floor physiography differs from that
described, for example, by Keene & Harris (1995)
from the Torres Strait of north-east Australia and
Longhitano & Nemec (2005) from the Amantea
Basin of southern Italy. In the latter case, the strait-
wide mounded complex of tidal dunes formed at
a greater water depth and developed considerable
lateral relief, with no wave influence and with the
dune cross-set thicknesses reaching 20 m and
showing recognisable lateral variation. A pro-
nounced sea floor relief and lack of wave impact
may thus be critical for a significant lateral differ-
entiation of dune heights within their subaqueous
fields such as tidal sand ridges.
()
(
nx x
n
>
)
EF x
=
i
i
(4)
i
In contrast to a conventional cumulative plot, the
exceedence frequency plot is 'decumulative' -
made by the back-stripping of data frequency for
successively higher
x
-values instead of piling them
up by accumulation. An exceedence frequency
0.000
1
-1.000
0.1
-2.000
0.01
-3.000
0.001
log
EF
(
x
) = 3.97 - 2.78 log
x
n = 666;
R
2
= 98.5%
(
SL
= 0.05%)
-4.000
0.0001
-5.000
0.00001
-6.000
0.00000
0.300 0.800
Measured max. thickness
x
max
= 250 cm
1. 300
1. 800 2.300
2.800
3.300
Estimation of cross-set volumes
Possible 'absolute' max. thickness
x
max
= 840 cm
Cross-set thickness logarithm (log
x
)
The dune cross-strata sets form natural deposi-
tional units, or beds, and are considered further as
such here. Their thickness measurements from
the individual cores (Fig. 8) have been analysed
by statistical methods, with a special focus on the
thickness frequency distribution and its implica-
tions for reservoir characterisation.
Cross-set thickness distribution
— The statistical
character of a dataset frequency distribution is most
Fig. 15.
The exceedence frequency plot of measured dune
cross-set thicknesses (bulk dataset from all six well cores)
matched by a lognormal distribution curve. Letter symbols:
R
2
- coefficient of determination (regression curve goodness
of fit);
n
- number of data;
SL
= 0.05% is the significance
level and means 99.95% confidence for the lognormal
trend. An extension of the lognormal curve to the lower
frequency limit suggests the possible occurrence of rare
cross-sets up to 840 cm thick.
Search WWH ::
Custom Search