574
angles. The incidence angles vary from
18° in the near range to 59° in the far
range.
method
makes full use of the multiple viewing
The soil moisture retrieval
capabilities of the ERS Scatterometer
(Wagner et al., 1998, Wagner, 1998).
Using the fore- and aft-beam antennas
the noise of c" due to instrument
noise, speckle, and large scale
terrain featurés can be estimated. The
incidence angle dependency of 0° can
studied
defined as the first derivative of oc?
using the slope which is
with respect to the incidence angle
and which can be calculated for each
Because bare
backscatter "triplet".
soil surfaces and vegetation are
characterised by a distinctly
different
the slope provides useful information
incidence angle behaviour
about vegetation conditions. Knowing
the spatial and temporal variability
of the slope the effects of vegetation
senescence on the
intensity a” can be
growth and
backscatter
accounted “for in the soil moisture
retrieval algorithm.
The fact that several years of ERS
Scatterometer data are available is
essential for the applied method. The
retrieval algorithm is in principle a
change detection method where the g^
measurements are compared to the
highest
recorded. As a
lowest and values ever
result a relative
measure of the surface soil moisture
content ranging from 0 to 1 (0.$ to
100 $) is obtained denoted by ©,. As a
working hypothesis it might be assumed
that the lowest 9^ values represent a
completely dry soil surface and the
highest c?
surface as might be encountered during
values a saturated soil
or shortly after a rainfall. In this
case ©, can be identified with the
degree of saturation of the soil
surface layer. This working hypothesis
might not be applicable to all regions
of the
investigated on a case to case basis.
world and needs to be
In the presented cases, the hypothesis
appears to be reasonable over the
Iberian Peninsula and Hungary. For
Mali where the landscape is
characterised by gently undulating
sandy soils saturation of the soil
surface layer may not be achieved.
This needs to investigated in more
detail in future studies.
The comparison of ERS Scatterometer
derived Os with rainfall
demonstrates the reliability of the
series
SOil moisture estimates. In Figure 1
Os is compared to rainfall
observations recorded at the synoptic
station Beja in the south of Portugal.
Peaks of ©, occur during or shortly
after rainfall events as e.g. in
November / December 1995. Some smaller
rainfall events (e.g. end of March
1994) are not reflected in the O09,
series because of lack of data. The
summer months are generally very dry
resulting in low ©; values. In 1995
the ERS Scatterometer shows that the
soil surface was completely dry from
April to October indicating a severe
drought. If a rainfall is followed by
a longer dry period an quasi
exponential drying process is observed
in the ©; “series (e.g. in January
1993, March:;1994 .and.March 1995). This
behaviour can be explained by the
redistribution of the water in the
wetted
relatively dry deeper layers (Hillel,
1980): "At first the decrease of soil
wetness in the initially wetted zone
surface layer into the
can be expected to occur more rapidly
during the redistribution of moisture
in profiles which had been subject to
shallow wetting than in the internal
drainage of profiles which had been
wetted
however, the redistribution process
deeply. Sooner or later,
"spends itself out" so to speak, and
the flux slows down for two reasons:
(1) the suction gradient between the
wet and dry zones decreases as the
former loses, and the latter gains
moisture; (2) as the initially wetted
zone quickly desorbs, its hydraulic
conductivity decreases
correspondingly”
3. DISCUSSION
Monthly soil moisture maps of the
Iberian Peninsula, Mali, and Hungary
are presented. The maps were produced
by calculating the monthly average
Intemational Archives of Photogrammetry and Remote Sensing. Vol. XXXII, Part 7, Budapest, 1998
