Geology Reference
In-Depth Information
The physical and geometrical characteristics of the
snowpack include snow density, snow grain size, snow
wetness, and ice layering and crusts within the snowpack.
Densification of snowpack increases its permittivity,
hence suppresses microwave emission from the underly-
ing sea ice [ Pullianen and Hallikainen , 2001]. Cycles of
melting and refreezing cause changes in snow grain
size, shape, and liquid water content in the snowpack.
Microwave brightness temperature increases nonmono-
tonically with snow grain size [ Fuhrhop et  al ., 1998].
Mätzler et al . [1984] and Mätzler [1994] elaborated on the
effects of different types of grains and layering of dry
snow on microwave scattering. Refreezing snow also
leads to the formation of ice layering (within the snow-
pack) and ice crust (near the surface of the snow base).
This causes significant scattering and hence a decrease of
T b especially at lower frequencies (≤37 GHz) [ Comiso
et  al ., 1997]. This effect is more pronounced for the
horizontal polarization at high incidence angles due to
the lower penetration depth and stronger snow layering
effects [ Hallikainen , 1989]. Tonboe et al . [2003] found also
that the refreezing of liquid water in the snow decreases
attenuation and increases scattering of microwave emis-
sion, leading to an increase in T b . Strum et  al . [2006]
presented detailed characterization of the physical prop-
erties of snow on sea ice in Barrow, Alaska. Based on the
measurements from 118 snow pits, they found that the
snowpack consisted primarily of layers of depth hoar
overlain by wind slabs. Fine‐grained and thin ice layers
were also observed in a few locations. The depth hoar
grain size distribution was highly mixed, consisting of
both large grains (2-10 mm) and small grains (0.1-0.3 mm).
Snow depth could also be related to ice surface roughness.
Strum et al . [2006] utilized this relation to generate approx-
imate maps of snow depth by identifying three ice types in
aerial photographs in the Alaskan Arctic: smooth, rough,
and moderately deformed ice. The snow depth was then
inferred from depth statistics related to these three types.
In general, snow depth algorithms using passive micro-
wave observations can be improved if snow properties,
especially grain size and layering, are made available coin-
cidently with the observations. In the meantime, because
of the uncertainties in snow grain size and density due to
sporadic weather effects, the snow operational products
from AMSR‐E are generated at 5 day averages.
The second set or factors that affect the snow structure
and its microwave emission encompass properties of the
underlying ice surface. This includes surface roughness,
salinity, the presence of a slush layer at the snow‐ice inter-
face, and apparently the type of underlying ice. These fac-
tors determine the initial amount of emitted radiation
that is subsequently scattered by the snow layer. By tak-
ing into account the sea ice emissivity, more accurate
snow depth retrieval algorithms can be developed.
Retrieval of snow depth on MY ice is difficult because of
the ambiguity between the emitted signal from the ice
surface and the snow cover. Markus et al . [2006] flagged
MY ice from the calculation of snow depth in the Arctic
using AMSR‐E data. Similarly, the operational snow
depth product from MODIS is generated only for the
seasonal sea ice zones in the Arctic.
The question of the snow depth retrieval from passive
microwave observations is linked to the understanding of
the emitted microwave signal in terms of the above‐men-
tioned factors. An efficient tool to achieve this under-
standing is modeling microwave emission from
snow‐covered ice under different conditions. Two models
are currently available, both evolved from an initiative by
ESA in 1995 to study snow over vegetation areas. The
first is a single‐layer model developed at the Helsinki
University of Technology (HUT) and is called the HUT
model [ Pulliainen et  al ., 1999]. This is a semi‐empirical,
zero‐order scattering model suitable for rapid emission
computation. The second is the Microwave Emission
Model of Layered Snowpacks (MEMLS), developed at
the University of Bern [ Wiesmann and Mätzler , 1999].
This is a multiple‐scattering model allowing many layers
to simulate snow cover emission with most of the known
physical parameters and processes. Although MEMLS
has been developed for snow on land, a version is used to
study sea ice emissivity by suitable scaling of the dielec-
tric and geometrical parameters [ Wiesmann and Mätzler ,
1999]. MEMLS is based on a radiative transfer model
and uses a correlation‐function approach to quantify
snow grains and their scattering mechanisms that include
(1) refraction and radiation trapping by total internal
reflection, (2) a combination of coherent and incoherent
superposition of interface reflections, and (3) multiple
scattering both by stratification and by snow grains.
Powell et  al . [2006] used the MEMLS model to study
the effect of snow on microwave emission and conse-
quently to validate the correlation between the gradient
ratio and snow depth, which was used to develop the
model of equation (10.103). They presented results from a
comparison of microwave brightness temperatures meas-
ured by the PSR instrument over the Elson Lagoon and
the Beaufort Sea along with simulations using MEMLS.
Input parameters to the model were obtained from ground
measurements of snowpack properties that included the
number of layers, thickness, density, surface and snow‐ice
temperatures, and prevailing grain size of each layer (con-
verted to correlation length). The study shows also that
the increase of correlation length of grain size affects the
emitted microwave radiation significantly and causes the
GR 37 V 19 V to become more negative (this is the advantage
of using the modeling approach to study the impact of
those physical factors on the emitted radiation). However,
the densification of a snowpack causes this parameter to
Search WWH ::




Custom Search