1036 
15m 
Table 1. Specif 
scatterometer 
Ad/sin0 
Figure 6. Relation be 
tween range resolution 
Ad and ground resolution. 
tion in general have to be supplemented.by physical 
parameters which can only be determined experimen 
tally. Before discussing some examples in section 4 
the experimental techniques will be described first. 
In general the required information is collected by 
radar systems of special design, the so-called scat- 
terometers. The underlying principles are basically 
the same as those used in radar. The first system to be 
mentioned is based on the well-known pulse-modulation 
principle. A pulse-modulated radar system measures 
distances by measuring the time interval t between a 
transmitted pulse and its echo. Since the signal trav 
els two times the distance d between the radar and 
the (point) target we have d=ct/2, where c is the 
velocity of light. Two targets which lie in the same 
direction as seen from the radar can only be detected 
separately when their echos have at least a separation 
At= T where T is the pulse duration. From this it 
follows that the distance resolution: 
Ad = cAt/2 = ct/2. (5) 
To handle a pulse of duration X the system needs 
a bandwidth B = 1/x and therefore we may as well 
write: 
Ad = c/2B (6) 
To relate the point target resolution to that for 
extended targets like in remote sensing we conclude 
from fig. 6 that a distance Ad in the direction of 
pulse propagation corresponds with a distance Ad/sin9 
along the ground. 
This ground distance can be identified as one of the 
two dimensions of the resolution cell. The other 
dimension perpendicular to the one first found fol 
lows from the antenna beam width. 
The second radar principle to be introduced here 
is the FM-CW principle where FM stands for frequency 
modulation and CW for continious wave. In the FM-CW 
scatterometer the transmitter is swept periodically 
over a frequency band B e.g. by means of a sawtooth 
wave (fig. 7) while the amplitude is kept constant. 
The transmitted frequency is in the figure indicated 
by the solid line. With distance d between radar and 
target the reflected signal will again be delayed by 
Figure 8. Configuration for test plot measurements 
Figure 9. Fluctuations of the 
radar backscattering when a 
scatterometer moves along test 
plots. 
time t=2d/c. The frequency of this echo signal is 
given by the dotted line. In the receiver the instan 
taneous frequency difference between the two signals 
is generated and from this so-called beat frequency 
the distance can be calculated. It can be shown (Krul 
1981) that the distance resolution for the FM-CW 
radar also follows from eq.(6). 
Finally we should mention the radars taking advan 
tage of the so-called Doppler shift appearing when 
the target and the radar have a relative velocity. 
When a target is moving towards the radar with a con 
stant velocity v m/s there will exist a frequency 
difference 2v/A between the transmitted and the re 
ceived signals. This frequency difference is called 
the Doppler frequency. The use of a Doppler scattero 
meter for extended target measurements requires spe 
cial measures to ensure proper resolution cells. 
Having ontlined briefly the principles used for 
scatterometry in relation to extended targets, we 
will describe the instruments developed for the re 
mote sensing program in The Netherlands. 
As was pointed out already in the introduction it 
was decided in 1973 to start the development of a 
scatterometer at 3 cm wavelength. The system was 
operated according to the FM-CW principle and was 
mounted on a 10m mobile elevator. The basic idea 
was that during the measurements the carriage was 
moving on rails along a series of test plots (fig. 8). 
During its travel the receiver output signal was re 
corded continuously . A sample of such a recording is 
reproduced in fig. 9, it clearly shows the different 
mean levels associated with the different crop types, 
on the test plots. 
For subsequent travels in general different inci 
dence angles and/or polarization states were chosen. 
The 3 cm scatterometer, the characteristics of which 
are summarized in table 1, was used (with slight modi 
fications) for crop measurements up to and including 
the growing season of 1981. Since it was to be expec- 
Frequency 
Frequency sweep 
Modulation wave 
Modulation freq 
Output power 
polarization 
incidence angle 
Range 
Antenna 
Recording 
Table 2. Speci 
based scatterom 
Frequency 
Frequency sweep 
Modulation wave 
Modulation freqi 
Output power 
Polarization 
Incidence angle 
Range 
Antenna 
Recording 
Table 3. Specif 
airborne scatte 
Frequencies 
Modulation type 
Pulse rep. freq 
Pulse’width 
Output power 
Polarization 
Incidence angle 
Range 
Antenna 
Data acquisitic 
ted that the ra 
only depends on 
of the electron: 
cy it seemed in 
quency. For a n 
chosen in the 8 
The specifics 
table 2. Only d 
1981 this 8mm s 
tic. measurement 
Although the 
to be very usef 
microwaves by v 
the first place 
very limited, t 
the a°-values f 
type (ecologica 
ondly the illun 
groundbased mea 
small, which sc 
scatter values 
systems. In the 
difficult to bu 
uencies mainly 
straints. Final 
most by definit