Major modifi- 
t to maintain 
na scanning 
t. 
Table 1. MSR 
| and 31.4 GHz 
10 msec and 
. temperature 
MSR is consisted 
bration sources. 
| load, while 
. would be pro- 
Radiometric 
cribed in the 
the input bright- 
of data from 
rce is a deep 
| temperature 
ation (1) 
switch 
itch 
d for loss and 
tric correction 
with loss L is 
cond term is the 
rce in the field 
    
  
experiment, cold calibration temperature T and corresponding digital counts N 
have to be estimated when hot calibration temperature T. and digital counts N 
are given. In order to make the estimation possible, rélationship between 
(T,,N )Jand (T, N ) was obtained through a thermal test prior to the ex- 
perîment. In the thermal test, radiometer was shrouded in a thermal blancket 
and instrument temperature was varied stepwisely. A cold reference target held 
at a constant temperature was placed in front of the sky horn and T, and N 
were obtained. s 
4. Method of observation 
MSR was installed in a frame with capability of adjustable elevation angle 
with handle (Fig.3). Polarization is also adjustable by rotating the radiometer 
about its electrical axis thus it is possible to measure brightness temperature 
of snow surface for incident angle ranging from 0 to 60 in vertical and horizon- 
tal polarization. 
The functional relationship of total observation system is indicated block- 
diagrmatically in Fig. 4. Observed brightness temperature and calibration 
source temperature are recorded onto an magnetic tape along with auxilary data 
such as annotation data and physical temperature of the components. 
The experiment was conducted from January 18 to 30, 1982. During the 
experiment, there were several snowfalls increasing snow depth from 55 cm to 90 
cm. Air temperature and snow surface temperature in the daytime were in the 
range of -2.5 C to -12 C and snow density at the surface varied about 0.06 g/cm 
to 0.13 g/cm”. The parameters of experiment are indicated in Table 2. 
During the observation with MSR, the ground truth data indicated in Table 3 were 
also obtained. 
3 
5. Results 
(1) Incident angle dependence 
During the experiment of January 18 to 30, 1982, snow surface temperature 
varied from -2.5 C to -12 C. So it can be assumed that there existed no liquid 
water and dry snow condition was prevalent. Typical data set of incident angle 
dependence is shown in Fig. 5.for horizontal and vertical polarization at the 
two frequencies. Measurement was made on January 20 when the minimum density 
(0.06 g/cm”) during the experiment was observed. 
In comparison with predecessors'results (Hofer et al., 1978, Tiuri et al.; 
1980), considerably less brightness temperature by 40 to 50 K was observed. 
This is considered due to dry snow condition where volume scattering is dominant 
and brightness temperature darkening effect takes place as stated by England (1975). 
Larger brightness temperature are observed fro 23 GHz than 31 GHz. The 
brightness temperature decreases with increasing incident angle in a similar 
manner for the two frequencies. The decreae is larger for horizontal polarization 
than vertical polarization. Temperature difference due to different polarization 
is larger for 23 GHz than 31 GHz. 
(2) Polarization angle dependence 
In MSR observation, a polarization plane rotates with respect to the ground 
polarization plane when antenna rotates. Since MSR operates in single polari- 
zation for the two frequencies, i.e. horizontal polarization for 23.8 GHz and