183
Symposium on Remote Sensing for Resources Development and Environmental Management / Enschede / August 1986
Relating L-band scatterometer data with soil moisture
content and roughness
P.J.F.Swart
Delft University of Technology, Netherlands
ABSTRACT: Preliminary results are reported of the analysis of L-band airborne radar backscatter data from lar
ge homogeneous agricultural fields. Examples are given of the calculated normalized radar cross section versus
incidence angle. The main theme of discussion will be this angular dependency and its relation with measured
soil moisture content and roughness.
RESUME: Dans cet article-ci les résultats préliminaires de l'analyse du signal rétrodiffusé d'un scattérométre
aéronautique functionnant àl.2GHz sont discutées. La région illuminée par le radar se compose des champs
d'agriculture homogènes de plus que septante hectares. Le thème principal est l'influence de la rugosité et
l'humidité du sol sur la dépendence du coefficient de rétrodiffusion en fonction de l'angle d'incidence.
1 INTRODUCTION
Within the framework of the Shuttle Imaging Radar-B
(SIR-B) experiment in October 1984 radar backscatter
data were gathered with the airborne scatterometer
DUTSCAT (Delft University of Technology Scatterometer)
working in L-band (1.2 GHz). The test area with large
sized agricultural fields lies in the southern part
of the Flevopolders in The Netherlands. The flights
with the scatterometer over these fields were perfor
med by the National Aerospace Laboratory (NLR). The
simultaneous collection of ground data was carried
out by the Soil Survey Institute in cooperation with
the Wageningen Agricultural University, both repre
sented in the ROVE Working Group Soils. The objectives
of the experiment were:
1. Calibration of the L-band radar on board of the
Space Shuttle.
2. Testing the inverse use of a scatter model for
bare soil.
Due to a number of technical problems the Shuttle
radar failed to produce data for our test area. Never
theless scatterometer and ground data as well as
flight parameters and video recordings were obtained.
This meant that the intended modelling could still
take place.
In the following the theoretical aspects concerning
the radar measurements will be described. After this
the processing of the acquired scatterometer data is
outlined. Some preliminary results of the angular de
pendency of calculated normalized radar cross sections
will be shown. Finally their relation with measured
soil moisture content and roughness is discussed.
2 MEASUREMENT DESCRIPTION
The DUTSCAT scatterometer is installed in the Beech-
craft Queen Air research aircraft of the NLR. It is
a well calibrated pulse type radar which is used for
accurate measurements of land and sea. A scatterometer
measures the power of scattered electromagnetic waves
quantitatively. Although any amplitude-calibrated ra
dar can be a scatterometer the most satisfactory mea
surements are made with radars designed as scattero-
meters. A technical description of the DUTSCAT system
can be found in literature (Attema e.a. 1984).
One of the most important properties of a scattero
meter is that in order to improve radiometric accuracy
a trade-off has been made against geometric resolution.
To allow an explanation of how this applies to DUTSCAT
we proceed with a synopsis of the underlying theory.
2.1 Theoretical aspects
In a pulse radar spatial discrimination between sig
nals received from different parts of the by the an
tenna illuminated area is achieved by measuring time
differences associated with different distances as
seen from the radar. The resulting so called slant-
range resolution is related to the time the radar
needs to create a pulse, the pulse duration T. This
relation is given by:
Ar = cT/2 (1)
with c the velocity of light (Krul 1986). We see in
fig. 1 that the resolution distance on the ground
perpendicular to flight direction follows as:
Ay = cT/2sin0 (2)
where 0 is the angle of incidence. For small incidence
angles this across-track ground resolution becomes
very large and ultimately the situation is reached
where the resolution is determined by the half-power
two way (or effective) antenna beamwidth in elevation.
For the along-track direction the resolution can be
approximated by the arc length corresponding to the
azimuthal effective beamwidth 0^:
Ax = B^R = B^h/cosB (3)
It follows that the resolution in x-direction degrad
es with increasing distance and/or incidence angle.
Figure 1. Pulse-type radar geometry.
