The Garn Formation (Bajocian-Bathonian) in the Kristin Field, Halten Terrace: its origin, facies architecture and primary heterogeneity model - From Depositional Systems to Sedimentary Successions on the Norwegian Continental Margin - page 533

1992), with reference to the core porosity and
permeability data (Fig. 13).
Sandstone porosities are known to be normally
distributed (Krumbein & Graybill, 1965). The
arithmetic mean and standard deviation of porosity
for the individual facies have thus been calculated
directly from the laboratory datasets. Sandstone
permeabilities, in contrast, are log-normally dis-
tributed (Krumbein & Graybill, 1965; Rosvoll, 1989)
and their corresponding statistics have been calcu-
lated from log-transformed, normalised datasets
(see formulae in Wilks, 1995, p. 95). Assuming that
the standard systematic core-plug measurements of
porosity and permeability represent adequately
variation of the two parameters in the sandstone
facies (Fig. 13), the general meaning of the cal-
culated statistics for a reservoir model would
be as follows: ~70 vol.% of a particular facies has
the parameter values in the range defined by the
mean ± 1 standard deviation; ~95 vol.% has values
in the range defined by the mean ± 2 standard devi-
ations; and ~ 100 vol.% has values in the range
defined by the mean ± 3 standard deviations. The
frequency distribution of each parameter for a
given facies volume would thus be fully described
by the two calculated statistics (Fig. 13).
Planar parallel stratification (facies S PS ) is not
considered to be a major source of reservoir hetero-
geneity, as the strata are not micaceous and gener-
ally lack strong textural variability. The sandstones
of facies S PS have a mean porosity of 13%, with a
standard deviation of 3% and a mean horizontal
permeability of 1.51 mD with a standard deviation
of 3.50 mD (Fig. 13A). Permeability in parallel-
stratified sandstones varies on the centimetre
scale, from one thin bundle of strata to another
(Rosvoll, 1989) and tends to be significantly higher
in bed-parallel than in bed-normal direction.
The ratio of vertical to horizontal permeability is
mainly in the range of 0.6 to 0.8. The main
permeability baffles are the truncation surfaces
separating strata sets, because these surfaces —
formed by episodic storms — are broad subtle
angular unconformities accompanied by abrupt
textural changes. There may be little difference in
the capillary properties of sandstone pores in the
direction parallel and perpendicular to strata but
the displacement of liquid hydrocarbon will
generally be more effective by bed-parallel flow
(Corbett, 1992). The arithmetic mean may be
adequate for both bed-parallel and bed-normal
permeability. The use of a harmonic mean, as
conventionally applied for bed-normal flow, leads

to a considerable overestimation of fractional oil
recovery (Corbett, 1992) and thus may also overes-
timate the flow of condensate liquid.
Ripple cross-lamination (facies S RL ) generally
involves finer-grained sand and thinner strata,
which renders permeability low, with mud flasers
and drapes acting as small-scale major baffles.
The porosity of facies S RL is similar to that of the
previous facies but its permeability is considera-
bly lower, with a mean of 0.92 mD and a standard
deviation of 1.99 mD (Fig. 13B). The ratio of verti-
cal to horizontal permeability is mainly in the
range of 0.4 to 0.7. ANOVA tests of permeability
probe datasets from similar sandstone facies indi-
cate that most variation (65% to 68%) occurs
within, rather than between, cross-lamina sets
and that the percentage of cross-set internal
variance may be as high as 82% to 88% in the
presence of mud drapes and flasers (Rosvoll,
1989). Most of the pore gas will be recovered but
the amount of condensate initially entrapped
by water capillarity may be significant due to the
volumetric percentage of this sandstone facies
(Fig. 12, Table 2).
Hummocky and swaley stratifications (facies S HS
and S SS ) are expected to have a relatively high
degree of internal heterogeneity, as they form sets
of fining-upwards strata that tend to be slightly
micaceous near the top and are often capped with
wave-ripple cross-lamination. The hummocky
and swaley strata sets have scoop-shaped ero-
sional bases and these angular unconformities are
commonly marked by an abrupt textural change.
The mean porosities of facies S HS and S SS are 12%
and 13%, respectively, both with a standard devia-
tion of 3% (Fig. 13C). The mean horizontal perme-
ability in facies S HS is 1.40 mD, with a standard
deviation of 4.36 mD, but is only 0.68 mD in facies
S SS , with a low standard deviation of 0.81 mD
(Fig. 13C). The permeability in either type of strata
set is expected to decrease upwards, which may
have little effect on the recovery of gas but may
affect the flow of liquid (see oil-recovery tests in
Corbett, 1992). However, the two facies are volu-
metrically subordinate (Fig. 12, Table 2) and hence
of minor significance to the bulk recovery.
Massive sandstones (facies S M ) are volumetri-
cally insignificant and represented by only one
core-plug sample in the dataset, with a porosity of
12% and a horizontal permeability of only 0.32 mD
(Fig. 13C). The parameters are lower than in the
underlying sandstones of similar grain size (facies
S CS or S PS ), which supports the notion of sand