can be used for angle discrimination within the illu 
minated spot. 
A video camera attached to the supporting struc 
ture of the antenna registrates the measured objects. 
A time indication added to the video images enables 
the data interpreter to link the desired objects 
with radar and flight data. The specifications of the 
DUTSCAT are listed in table 2. 
Table 2. Specifications of DUTSCAT (multi-band). 
Frequencies 
Antenna 
Transmitted power 
Operating range 
Pulse length 
PRF 
Polarization 
Sample frequency 
Range resolution 
1.2, 3.2, 5.3, 9.65, 
13.7 and 17.25 Ghz 
0.9 m parabolic dish 
250 mW 
50 - 1920 meters 
100 ns 
78.125 kHz 
HH or VV 
20 MHz (7.5 m), 
8 bits I & Q 
15 m 
The forest stands of interest have been measured 3 
or 4 times with the DUTSCAT for each angle and/or 
polarization. For the relatively small stands and for 
the high incidence angles this repetition was found 
to be necessary. This can be explained as follows. 
The distance r between sensor and object and the 
ground range distance y (figure 1) will fluctuate 
around designed values. The effect of small roll 
movements and height variations of the platform can 
easily be quantified. Since y = h. tan (0^) 
h 
dy = d0£ + tan0- ,dh (1) 
cos 2 0£ 
This formula shows that especially at high incidence 
angles variations in dh en d0£ will have a strong 
effect. This is illustrated by the following numeri 
cal example: Suppose the height of flight is 200 
meter and angle of incidence is 60 degrees. Realistic 
values for height and roll variation are dh= + 10 
meter and d0^ = +2 degrees. This results in 
y = 346.4 meter and dy = + 48.6 meter. 
The variation in dy is troublesome when small stands 
have to be measured. If navigation is perfect, 
variations in height and roll angle will still cause 
the illuminated spot to sweep across the stand and to 
cross borders occasionally. Therefore only the lar 
gest forest stands (exceeding 200 m in ground range) 
have been selected for the DUTSCAT campaigns. 
For research the multiband scatterometer has some 
major advantages over the SLAR. Gamma values can be 
measured simultaneously in six frequencies, a choice 
can be made between vertical and horizontal like 
polarization and the system is accurately calibrated 
both internally and externally. The major advantage 
of the SLAR is its imaging capability. Spatial 
patterns and differences in image texture can be per 
ceived. 
Another difference between SLAR and scatterometer 
is not apparent immediately. It follows from diffe 
rences in measurement geometry. The SLAR usually is 
operated from larger altitudes and the across track 
beam width is considerably larger making the across 
track illuminated spot width in the order of seve 
ral kilometers whereas this figure is in the order of 
several tens of meters for the scatterometer (figure 
2). As a result contributions of scatterers from 
different horizontal layers in the forest arrive 
somewhat separated in time at the receiver (section 
4). This effect enables the data interpreter to 
distinguish sources of scattering (sections 4 and 5) 
but at the other hand makes 'standard' pre 
processing as used for non-forested areas question 
able (sections 4 and 6). 
Figure 1. Measurement geometry with height of flight 
(h), distance sensor - target (r), ground range dis 
tance (y) and angle of incidence (0f). 
Figure 2. The relatively small beam width and low 
height of flight of the scatterometer (DUTSCAT) ma 
kes that contributions of scatterers from different 
horizontal layers in the forest are measured in dif 
ferent range cells. For the SLAR system in x every 
range cell all forest layers are (equally) repre 
sented. 
3. FOREST CANOPY ATTENUATION MEASUREMENTS USING 
CORNER REFLECTORS 
3.1 Corner reflector experiment 
During the summers of 1984 and 1985 measurements of 
forest canopy attenuation were conducted with the use 
of corner reflectors. The first experiment of this 
kind took place at the Roggebotzand site over stands 
of poplar and oak with the X-band SLAR. In 1985 these 
measurements were continued with the X-band 
and the L-band scatterometer but these have not been 
fully analysed yet. Three corner reflectors with an 2 
X-band raaar cross section (in free space) of 45 dBm