Geology Reference
In-Depth Information
and white image). The gray shade cutoff for the ice‐
brine partitioning in the cases of young and FY ice
was determined using the brine volume of the sample,
which was determined using equation (3.26). After
identifying the inclusions using standard image pro-
cessing software, a suite of parameters was calculated:
the number and size distribution of the inclusions,
their major and minor axes, orientation, and shape fac-
tor. A main finding from that study was the nearly per-
fect fit of the cross‐section area of inclusions to the
lognormal distribution. This was concluded from anal-
ysis of 40 thin sections encompassing a wide range of
ice types (with granular and columnar crystalline struc-
ture) and including cases with brine pockets and air
bubbles. The lognormal form of the cumulative distri-
bution function (CDF) of the cross‐sectional area A of
the inclusions is given by
0.0
0.5
1. 0
1. 5
2.0
2.5
0.5
1
2
(ln
A
)
0.6
0.7 0.8
Dentritic spacing (mm)
0.9
1. 0
1.1
1. 2
CDF
()
A
erfc
(4.6)
2
Figure 4.53 Dependence of brine layer spacing on depth for
columnar‐grained sea ice in Resolute Bay, Canadian High
Arctic. The ice core from which the data were obtained was
extracted in May 1992. Open circles and squares indicate field
and laboratory measurements, respectively, while solid trian-
gles are averaging brine spacing [ Sinha and Zhan, 1996].
where μ and σ are the mean and standard deviation of
ln  A . The probability density function (PDF), which is
the first derivative of the CDF, is given by
1
1
(ln
A
)
2
(4.7)
PDF(
A
)
exp
A
2
2
2
2
1. 0
Perovich and Gow [1996] pointed out that the lognor-
mal distribution is common for natural parameters
when random processes contribute multiplicatively
rather than additively to the parameter (the latter case
leads to a normal distribution). They also indicated that
the lognormal distribution has been used to describe
pressure ridge keels ( Davis and Wadhams, 1995) and
grain coarsening in snow [ Colbeck, 1987]. An example
of the observed CDF and the derived PDF of the brine
inclusion area from a horizontal section of columnar
young ice is presented in Figure 4.55. The deviation of
the measured area from the lognormal distribution at
the lower end of the area scale is due to the limitation
of the software in measuring very small areas as indi-
cated before. The upper area limit of 0.1 mm 2 in the
figure is also smaller than the findings in other studies
[ Cole and Shapiro, 1998; Light et al., 2003].
A more recent study by Light et al. [2003] used an auto-
mated imaging system for observing ice microstructure in
vertical thin sections prepared from land‐fast ice near
Point Barrow, Alaska, in May 1994. The system allows
detection of brine inclusions as small as 0.003 mm. This is
much smaller than the minimum dimension of approxi-
mately 0.05 mm presented in Perovich and Gow [1996]. This
0.9
0.8
λ 1 = 1. 11 -0.093 S
0.7
0.6
0.5
1
2
4
5
6
0
3
Salinity ( S in ‰)
Figure 4.54 Dependence of brine layer spacing on salinity for
mature columnar‐grained sea ice in Resolute Bay in May 1992
[ Sinha and Zhan, 1996].
types (young, FY, and MY). They used an image pro-
cessing system to partition images of thin sections of
sea ice into constitutive components of ice, brine (in
FY ice), and air (in MY ice). This was achieved by
exploiting differences in gray shade between ice and its
inclusions (inclusions are the darkest features in a black
 
Search WWH ::




Custom Search