Geology Reference
In-Depth Information
as well as the results obtained 12 h later when the rubble
pieces had frozen to the ice surface. Data from the verti-
cal polarization are easier to interpret because the emis-
sivity of the smooth (undisturbed) surface remains nearly
constant at all frequencies. At 6 GHz, the emissivity from
vertical polarization is the same after adding the rough-
ness disks. This shows how the 2 cm thick disks appear
smooth to the 6 GHz emission (the equivalent wavelength
is almost 5 cm). The presence of surface roughness
decreases the emissivity at all higher frequencies (when
scattering takes place). Freezing of the rubble has virtu-
ally no effect on the spectrum of the emissivity.
The situation of the horizontal polarization is more
complex. The undisturbed ice surface shows a distinct
minimum near 8 GHz followed by a gradual increase to a
maximum near 18 GHz, and then it decreases at a slight
rate.
Grenfell et al
. [1988] explained the minimum emissiv-
ity as being the effect of a first‐order fringe pair of the
surface that were captured at lower frequencies. On the
other hand, the higher order fringes that should have
been captured at higher frequencies were not observed
due to attenuation and scattering in the frost layer that
covered the surface at the time when the rubble layer was
introduced. When the rubble ice layer was added the
emissivity at low frequencies rose by about 0.15 at low
frequencies and then decreased monotonically at higher
frequencies. Figure 8.37 shows also that surface roughness
triggers more emission at vertical polarization than at
horizontal polarization. The undisturbed surface also has
higher emissivity in the vertical polarization. Data pre-
sented in this section support the notion that horizontally
polarized emission is more affected by volume scattering
while vertically polarized emission is more affected by
surface scattering. Modeling the emission is the best tool
to explore these relationships. Modeling microwave emis-
sion from sea ice has not advanced enough [
Heygster
et al
., 2006;
Mathew et al
., 2008], although modeling the
emissivity from ocean surface have matured enough for
applications [
Wentz
, 1983].
The addition of dry snow reduces the emissivity of sea
ice due to increased scattering in the snow volume. This
was shown in
Grenfell and Comiso
[1986] based on meas-
urements from simulated sea ice grown in an outdoor
tank in the CRREL. Results are presented in Figure 8.38
from measurements before and after a snowfall of about
4.5 cm on ice with average thickness of 12 cm (data after
snowfall are not shown for the 90 GHz). The fresh snow
reduces the emission more in the horizontal polarization
than the vertical polarization as show in the figure. The
observed angular dependence of emissivity in Figure 8.38
is similar to field observations in the Beaufort Sea
reported in
Grenfell and Lohanick
, [1985].
Moisture in the snow causes higher absorption of
energy. Therefore wet snow becomes a good emitter (i.e.,
has higher emissivity). Data showing the effect of snow
wetness on the emissivity of the snow‐covered sea ice are
not readily available in the literature. However, the gen-
eral notion is that as snow acquires wetness, volume
scattering becomes insignificant and that leads to an
increase in emissivity until saturation is reached near
1.0. Researchers usually use this notion to interpret their
observations. This is the argument used, for example, to
1. 0
0.9
0.8
0.7
H polarization
V polarization
0.6
No rubble
25 mm rubble layer
Rubble frozen to ice
No rubble
25 mm rubble layer
Rubble frozen to ice
0.5
0.4
4
10
40
100
4
1
0
40
100
Frequency (GHz)
Frequency (GHz)
Figure 8.37
Emissivity spectra of 150 mm thick ice before and after addition of a 25 mm layer of ice rubble
[
Grenfell et al
., 1988, Figure 3, with permission from IEEE].
