In: Wagner W., Szekely, B. (eds.): ISPRS TC VII Symposium - 100 Years ISPRS, Vienna, Austria, July 5-7, 2010, IAPRS, Vol. XXXVIII, Part 7B 
179 
Figure 7: Example of a histogram of retrieved soil moisture con 
tent (mUretr) 
mV ret r [vol%] 
5°r - 
10 ■ 
1 1 1 1 1 
0 10 20 30 40 50 
mv me as [vol%] 
Figure 8: Retrieved (mu re tr) versus measured soil moisture 
(mvmeas), the error bars represent IQR/1.35 
retrieved histograms. Furthermore a Nash-Sutcliffe model effi 
ciency (Nash and Sutcliffe, 1970) of 0.63 was found, indicating 
the model predicts much better than the mean value of the ob 
servations. It is furthermore observed that the uncertainty on the 
retrieved soil moisture contents increases with the soil moisture 
content, which is in accordance with the observations of Verhoest 
et al. (2007). 
4 CONCLUSIONS 
This study presents a methodology that allows for the 
retrieval of soil moisture content and its uncertainty based on 
modeled roughness. Soil surface roughness, in terms of effective 
correlation length (Z e fr) is modeled based on its relationship with 
normalized backscatter coefficients (cr„). The uncertainty on the 
modeled correlation length / mo a is described by a t-distribution, 
which is then sampled following a Monte Carlo method. The 
randomly drawn values are propagated through the inversion of 
the IEM and a corresponding histogram of soil moisture contents 
is obtained. 
Results show that most of these histograms are skewed and non 
normal and that a representation of these histograms by means 
of the mean value and the standard deviation may lead to a dis 
torted view of the underlying distribution. This is particularly 
important when retrieved soil moisture content and correspond 
ing uncertainty (represented by the mean and standard deviation) 
are to be used in data assimilation schemes, such as the Ensemble 
Kalman filter, which rely on normality assumptions of the vari 
ables of interest. It would be better to apply the median value and 
converted IQR in the data assimilation framework. It is further 
more observed that the interquartile range changes with varying 
soil moisture conditions, larger interquartile ranges are obtained 
for higher soil moisture contents. 
Future research is required to test whether soil moisture content 
with a variable uncertainty has a large impact when used in a data 
assimilation framework. 
ACKNOWLEDGEMENTS 
This work has been performed in the framework of the STEREO 
II project SR/00/100 “Hydrasens”, financed by the Belgian Sci 
ence Policy. RADARSAT-1 data was provided by the Canadian 
Space Agency (CSA) project DRU-10-02. Spanish Government’s 
project CGL2007-63453/HID is also acknowledged. 
References 
Alvarez-Mozos, J., Casali, J., Gonzalez-Audicana, M. and Ver 
hoest, N. E. C., 2006. Assessment of the operational applica 
bility of RADARSAT-1 data for surface soil moisture estima 
tion. IEEE Transactions on Geoscience and Remote Sensing 
44,pp.913-924. 
Alvarez-Mozos, J., Gonzalez-Audicana, M., Casali, J. and 
Larranaga, A., 2008. Effective versus measured correlation 
length for radar-based surface soil moisture retrieval. Interna 
tional Journal of Remote Sensing 29, pp. 5397-5408. 
Attema, E. P. W. and Ulaby, F. T., 1978. Vegetation modeled as a 
water cloud. Radio Science 13, pp. 357-364. 
Baghdadi, N., Gherboudj, I., Zribi, M., Sahebi, M., King, C. 
and Bonn, F., 2004. Semi-empirical calibration of the IEM 
backscattering model using radar images and moisture and 
roughness field measurements. International Journal of Re 
mote Sensing 25, pp. 3593-3623. 
Baghdadi, N., Holah, N. and Zribi, M., 2006. Calibration of the 
Integral Equation Model for SAR data in C-band and HH and 
W polarizations. International Journal of Remote Sensing 27, 
pp. 805-816. 
Baghdadi, N., King, C., Chanzy, A. and Wigneron, J. P., 2002. An 
empirical calibration of the Integral Equation Model based on 
SAR data, soil moisture and surface roughness measurement 
over bare soils. International Journal of Remote Sensing 23, 
pp. 4325-4340. 
Biftu, G. F. and Gan, T. Y., 1999. Retrieving soil moisture from 
Radarsat SAR data. Water Resources Research 35, pp. 1569- 
1579. 
Callens, M., Verhoest, N. E. C. and Davidson, M. W. J., 2006. 
Parameterization of tillage-induced single-scale soil roughness 
from 4-m profiles. IEEE Transactions on Geoscience and Re 
mote Sensing 44, pp. 878-888. 
Davidson, M. W. J., Le Toan, T., Mattia, F., Satalino, G., Man- 
ninen, T. and Borgeaud, M., 2000. On the characterization 
of agricultural soil roughness for radar remote sensing stud 
ies. IEEE Transactions on Geoscience and Remote Sensing 
38, pp. 630-640. 
Dierking, W., 1999. Quantitative roughness characterization of 
geological surfaces and implications for radar signature anal 
ysis. IEEE Transactions on Geoscience and Remote Sensing 
37,pp. 2397-2412.