Geology Reference
In-Depth Information
1. 0
1. 0
MY ice-
melt pond
V polarization
H polarization
Grease ice
0.8
0.8
NILAS
Light (10+ cm)
Gray (4-6 cm)
Dark (
~
3 cm)
0.6
0.6
MY ice-
hummock
V polarization
H polarization
0.4
0.4
0.2
0.2
1
10
100
1
10
100
Frequency (GHz)
Frequency (GHz)
Figure 8.33
Emissivity of new ice types: grease ice and three
stages of Nilas of three thickness ranges, plotted versus fre-
quency. Data were obtained from measurements of brightness
temperature at 50° incidence angle during experiments in the
Weddell and Bering Seas [
Eppler et al
. 1992, Figure 4‐3, with
permission from AGU].
Figure 8.34
Emissivity from the two surfaces of MY ice: hum-
mock and melt pond, as a function of frequency averaged from
measurements taken over a number of years at a range of Arctic
sites [
Grenfell
, 1992, Figure 4, with permission from AGU].
bubbly layer at the top surface of hummock ice is a highly
scattering medium, which causes emissivity to decrease.
This effect is enhanced at higher frequencies as the wave-
length approaches the typical dimensions of the air bub-
bles as mentioned before. On the other hand, melt pond
ice freezes from pond water at the ice surface is much
less porous and therefore gives rise to higher emissivity
[
Grenfell
, 1992]. In this case the emissivity approaches 1.0
at all frequencies >10 GHz as shown in Figure 8.34. If
thick snow covers both hummock and melt pond ice, the
difference in their emissivity is reduced. This complicates
the estimation of emissivity from MY ice. Figure 8.35
confirms the separability between hummock and melt
pond ice in the 2D space of the 18 and 37 GHz emissivity
(data obtained from the same sources as in Figure 8.34.
The emissions from the two frequencies are highly corre-
lated. This correlation is generally observed between fre-
quency channels at different polarization or polarization
channels at different frequencies, regardless of the type of
ice cover. Once again, the lower values correspond to data
from hummock ice.
An attempt to relate emissivity to surface conditions
of young ice and establish emissivity‐based surface cat-
egories is presented in
Shokr et al
. [2009]. Microwave
emission measurements from simulated thin sea ice (up
to 20 cm thick) in an outdoor tank were converted to
emissivity using equation (7.28) after substituting
T
b
with
T
b
down-ref
from equation (8.19). In addition to the
OW, emissivity was calculated for six surfaces developed
during the experiment: wet slush, wet bare ice, refrozen
slush, wet snow cover, dry snow cover, and dry bare ice.
Results are presented in Figure 8.36. Dry surfaces (bare
here means equal microwave emission in the horizontal
and vertical polarization). The first reason is the decrease
of the contribution of the polarized emission from the
underlying sea water to the total emission from the ice‐
water composition. Laboratory observations show that
ice becomes optically thick at 60 cm thickness for 10 GHz
and less than 10 mm thickness for 90 GHz [
Grenfell and
Comiso
, 1986]. Above that thickness the underlying water
no longer makes a contribution to the overall emissivity.
The second reason is the increase of volume scattering as
the ice thickens, which increases the depolarization of the
radiation.
Eppler et al
. [1992] pointed out that changes in
emissivity observed with the growth of the ice sheet can-
not be related solely to the ice thickness. That is because
the salinity profile and the overlain snow are also key fac-
tors in determining the emissivity.
Emissivity of MY ice is less complex than that of FY
and younger ice types. For one thing, the ice surface has
become much more stable in terms of its salinity (mostly
desalinated) and thermal state. Moreover, the overlain
snow is nearly saline free and is also stable during the cold
winter in the Arctic (where MY ice covers the central area
of the Arctic Ocean). One of the complexities associated
with determining the emissivity of MY ice is the differ-
ence between hummock and melt pond ice surfaces (see
the difference in physical properties between these two
entities in sections 2.5 and 4.4.3). Figure 8.34 shows that
the emissivity of the hummock surface is considerably
less than that of the melt pond surface, especially at
higher frequencies. Data were obtained from five field
experiments in the marginal ice zone in the 1980s. The
