nd
values
are
re
y and
at
once
ad at
ric
3 near
riza-
ariza-
, the
rfaces
cion
Dr
on-
is the
cies
ints of
>. multi
ler
res.
RA x
Calm water clearly lies on the Fresnel line at both frequencies.
Whereas the windy sea points depart from this line due to polari-
zation mixing on the tilted surfaces, new ice goes more or less
along the Fresnel line toward higher emissivities. A slight in-
crease in polarization. ey-eh, Point A) results from an anisotrop-
ic effective dielectric constant: The frazil ice consists of most-
ly horizontally extended platelets with diameters of about 1 cm
and a thickness of a fraction of 1 mm. Multiple layers of these
platelets form the soft and wet pancake-like clusters with a simi-
lar ratio of diameter to thickness as the single ice crystals.
This anisotropy results in larger dielectric constant in horizont-
al than vertical direction, which explains the observed departure
from the Fresnel line toward high polarization.
When the ice becomes densely packed, the surface roughens,
reducing the polarization (Points B to C). It is
note that snow-free, wetted multi year ice shows
vity as thick pancake ice.Thus the emissivity is a unique function
of wetness. For the very young ice the thickness dependence ap-
pears indirectly via the change of wetness with increasing ice
thickness. Very close to the black body point are the emissivities
of wet snow also shown in Figure 3. The wetness range of these
snow data is between less than 1$ and 10$ by volume. The emis-
sivities decrease with decreasing frequency and increasing wet-
ness, respectively.
thus
interesting to
the same emissi-
When plotted against thickness of the young ice, the emissivity
curve at 10.4.GHz appears as shown in Figure 4 (509 nadir angle,
horizontal and vertical polarization). A similar behaviour is
found at all frequencies. The emissivities show the monotonic in-
crease with thickness, the critical thickness being smaller at
vertical than at horizontal polarization.Both critical thicknesses
are slightly smaller than the critical heat exchange thickness.
The Points A, B, C on the thickness axis refer to data also shown
in other figures.
3. ACTIVE MICROWAVE SIGNATURES (at 10.4 GHz)
Figure 5 shows 3 angular scans of like-polarized backscattering
coefficients of small pancakes in water (A), densely packed pan-
cake ice (B) and thick pancake ice (C). The situation A shows a
pronounced difference between dj, and %yy supported again by the
anisotropy of the dielectric constant. Backscattering in this case
can be described by diffraction from the randomly absorbing sur-
face with a variance higher at vertical than at horizontal polari-
zation (Schanda, 1982). The large backscatter from the rough sur-
face of situation B is similar to the backscattering from multi
year ice. This ambiguity is a potential for confusion in radar
data. Thick pancake ice (C) shows smaller backscattering because
of the smaller roughness and wetness than in situation B. When the
ice is covered by wet snow, the backscattering decreases further
and approaches values near “= 0.01 (-20 dB).
Figure 4 shows the variation of the backscattering at 509 nadir
angle with increasing ice thickness. The vertical "error" bar at
zero thickness shows the range of f. for calm and windy (7m/s)
sea. The corresponding values for ¥, are slightly higher. When
ice forms, the sea roughness first decreases due to the damping
of the capillary waves. Therefore, Figure 4 shows first a decrease
of ¥. Then the inhomogeneity (Figure lc, Situation A) and much
761
m
1r. 3 ERE SS
5
+
à = = i.
RE