The accuracy of these data is better than 1% at 4.9, 10.4 and E
36 GHz, a non-linearity at 21 GHz shifts the 21 GHz.data. to. values zo
too high at low emissivities up to 3$, and the 94 GHz data are alc
accurate to 10%. With these uncertainties in mind, the figure CI:
shows the increase of emissivities with increasing frequency and ic
ice thickness. The increase starts faster at vertical than at ly
horizontal polarization. There is no indication of interference an
effects from superpositions of reflections at the surface and at pl
the ice-water interface, in agreement with the high dielectric m
loss found for young frazil ice (Vant et al. 1974). Th:
For thicker pancake ice the spectra converge to emissivities near al
0.95 independent of frequency and very similar at both polariza- Er.
tions. Whe
Figure 3 shows emissivities, vertical versus horizontal polariza- rec
tion at 4.9 and 36 GHz of wet ice and wet snow. In addition, the noi
theoretical range of emissivities for Fresnel reflecting surfaces Vii
is shown. of
pe:
th.
1 T T 1 I T of
sn
ey} 50° nadir angle six
ne:
Whe
8} CU:
ho:
fo
L Cr:
ve:
ar:
Thi
.6 |- in
8.
i Fi:
© windy sea , 7m/s Co
Ar e calm, open water gj pr:
—O— young sea ice an.
_ —o— wet, snow-free oi Ca
multi-year ice
za
fa
2 1 l 1 1 ye.
"To 2 4 .6 8 en 1 da
Figure 3: Emissivities, vertical versus horizontal polarization of
at 50° nadir angle. The solid line "Fresnel" shows values for ic
which the Fresnel formulae are valid with real dielectric con- an
stant (£21). The dashed line approaching this solid line is the
Fresnel curve for complex E= €- i €" with €' - 1 andt"»0. Fi
All intermediate dielectric constants have Fresnel emissivities an
in the narrow range between the two lines. Measured data points of ze:
calm and slightly rough sea, young sea ice, wetted snow-free multi se
year ice, and wet snow at 4.9 and 36 GHz are shown. ic
The letters A (small pancakes in between water), B (densely of
packed pancakes), and C (thick pancake ice) refer to the other of
Figures.
760
DESI — "m =
