475 
Microwave Radiometry of Snowpacks 
Christian Mätzler and Thomas Weise 
Institute of Applied Physics, University of Bern, Sidlerstr. 5, CH-3012 Bern 
Summary 
The successful application of passive microwave sensors to remote sensing requires signatures 
for the unambiguous inversion of the observable data to useful geophysical information. Due to 
the large number of object types, large variability and heterogeneity, the inversion of satellite 
data from land surfaces is delicate. The situation gets even more complex with the addition of a 
seasonal snowcover with variable transparency. We would like the sensors to be able to 
- discriminate snow-covered areas against any other surfaces, 
- determine the snow-cover fraction in mixed situations, 
- determine snow type, 
- determine snow properties such as depth, water equivalent, liquid-water content, density, 
grain size, snowpack stability etc. 
- estimate other physical properties such as ground temperature, soil moisture, presence of ve 
getation etc. 
A necessary, but not sufficient condition is a detectable sensitivity of the observables to the 
above properties. It has been a long-term task of the terrestrial remote sensing group of our In 
stitute to evaluate these sensitivities over a broad spectral range (1 to 100 GHz). The research 
has been concentrated on experimental studies with surface-based equipment The present state 
of knowledge over the frequency range from 4.9 to 94 GHz was recently analyzed by Matzler 
(1994). Most of these results were obtained with the multi-frequency radiometer system, PA 
MIR (Matzler, 1987). Relevant results will be presented. Ambiguities of observables on the 
one hand and unexplained variations on the other limit the application of these results as direct 
signatures. Nevertheless we found that the majority of snow types can be discriminated against 
the considered snow-free surfaces. Difficulties exist with fresh snow at low density where only 
the 94 GHz brightness temperatures indicate a sensitivity due to slight volume scattering. The 
earlier difficulty with wet snow (Matzler et al. 1982) has also found a solution; it is based pri 
marily on the spectral behaviour of the polarization difference, T^-T^, between the brightness 
temperatures at vertical and horizontal polarization. The results have also shown that the state 
(frozen or unfrozen, wet or dry) of the soil under a grass cover hardly affects the observable 
brightness temperatures, because of the screening effect by the vegetation. 
As illustrations of physical snow signatures we presented and discussed the time variations of 
brightness temperatures as observed during two different episodes at the alpine test site, 
Weissfluhjoch in Switzerland. 
The first example was an interference phenomenon observed during the night of June 18 to 19, 
1984 at 4.9 and 10.4 GHz, respectively, (Fig. 4.26 of Matzler, 1987) which can be explained 
by the superposition of reflections at the refrozen snow surface and at the down-moving inter 
face between refrozen and wet snow. The data allowed the determination of snow density