2. DATA COLLECTION 
2.1 Location and Site Description 
The experiments were conducted at the Soil 
Science Laboratory of the Institut National de la 
Recherche Agronomique (I.N.R.A.) of Montfavet 
(France) during July 1988. The soil was a clay loam 
with about 27 % of clay and 62 % of silt. 
The field (0.1 ha) was initially tilled with a rotary 
digging machine. Soil clods were arranged in an 
apparently random manner and resulted from soil break 
up tillage implements (no tools marks or tillage 
directions). The general soil surface was flat. 
Rainfall was simulated using a sprinkler irrigation 
equipment composed of an aluminium frame (20 m 
long) with regularly distributed sprinklers and 
supported by two pneumatic tyred wheels on each side. 
The irrigation system was moved across the 
experimental field on two tramlines located on each side 
of the field. Rainfall was applied during 4 successive 
periods to simulate five stages of slaking (including the 
initial stage). The rainfall intensity was approximately 
40 mmh' 1 . The cumulative rainfall precipitation was 
measured. 
At each stage of slaking, ground observations 
concurrent with remote sensing measurements were 
performed on the experimental field under wet and dry 
surface soil moisture conditions. Wet condition is 
obtained after completely surface draining (no puddles). 
Dry condition is reached by natural evaporation during 
two or three days. 
2.2 Remote sensed data. 
Table 1 : Spectral band of the SPOT 
simulation radiometer. 
Band 
Wavelength (nm) 
XS1 
500 - 590 
(a) 
XS2 
600 - 690 
(b) 
XS3 
790 - 900 
(c) 
(a) 
green band 
(b) 
red band 
(c) 
near infrared band 
This instrument has two heads, (l)the first is equipped 
with diffusing screens and measures irradiance. It was 
fixed at the bottom of the crane. (2)the second is 
equipped with a 12° field of view and measures field 
target radiance. It was fixed at the mobile platform. 
The plot position, in respect to viewing and 
illumination directions are seen in Figure 2. The 
radiometer is enable to simulate SPOT sattelite data 
taken in nadir viewing angle (0* from nadir) and 
oblique viewing angle, here 23° (from to nadir). The 
choosen angle of incidence was different but quite near 
the oblique viewing angle of the SPOT satellite (27‘ 
relative to nadir). One cloud free day, thirty 
radiometric measurements in the three spectral bands 
were recorded during the travel of the platform and 
averaged. Radiometric measurements were performed at 
three approximately constant solar illumination angle at 
6.00 am, 12.00 and 6.00 pm G.M.T. 
The two instruments were mounted an a mobile 
platform of a crane-boom. They were at about 19.70 m 
height above the soil surface. Remote sensing 
measurements were performed on about one third of the 
experimental field (Figure 1). 
2.2.1 Microwave measurements. We used the 
scatterometer RAMSES designed by the Centre 
National d'Etudes Spatiales (C.N.E.S.) of Toulouse 
(France). It is a frequency-modulated continuous wave, 
multifrequency and multipolarization system. 
Microwave measurements were made at 5 3 GHz (C 
band) and HH polarization. According to theoretical 
and experimental results (Ulaby et al., 1978; Ulaby et 
al., 1982), we used only small or large angles of 
incidence to increase the sensitivity of microwave 
measurements to roughness changes. The two angles of 
incidence were fixed at 0‘ (nadir j and 50' from nadir. 
The surface foot print varied from 3 m 2 at nadir to 15 
m 2 at 50*. Microwave measurements were expressed in 
term of the backscattermg coefficient (<r°) and reported 
in units of decibels. Spatial averaging of the <r° 
measurements was achieved by recording 200 <r° values 
during the travel of the platform along the crane-boom. 
2.2.2 Optical measurements. They were performed in 
three different spectral bands (Table 1.) using a SPOT 
simulation radiometer (Guyot et al, 1984). 
2.3 Ground measurements 
Depending on soil moisture conditions, ground 
observations of soil moisture, soil roughness and 
determination of stage of slaking were performed on the 
the field area not seen by the intruments (Figure 1). 
Ten gravimetric water content samples were collected 
approximately at the same location to measure soil 
moisture (Figure 1). They were taken from 0 to 2 cm 
depth. Dry bulk densities were measured at 3 cm depth 
using a gamma transmission probe (Bertuzzi et al., 
1987). They were extrapolated at the 0-2 cm soil layer 
to compute volumetric water contents by combining 
gravimetric water contents and dry bulk densities. 
An automated non-contact Laser profile meter was 
used to measure changes in soil surface roughness 
(Bertuzzi et al., 1990a). Each profile was 2 m long with 
1000 elevation data recorded at 2 mm sampling 
interval. The accuracy on elevation data was less than 
0.25 mm. After each rainfall, twelve profiles, regularly 
distributed on the experimental field were recorded 
always at the same location (Figure 1). Six profiles were 
recorded in the main longitudinal eastern -» western 
direction of the field, parallel to the tillage direction. 
The six other profiles were recorded in the nothern -> 
southern direction perpendicular to the tillage direction. 
At each stage of slaking, roughness profiles were used to 
compute two roughness indices. The first one is the 
most common standard deviation of height (s). The