|
Spatially distributed soil moisture is required for watershed applications such as drought and flood prediction, crop
irrigation scheduling, etc. In particular, an accurate assessment of the spatial and temporal variation of soil moisture is
necessary to improve the predictive capability of runoff models, and for improving and validating hydrological processes
forecasting. In recent years, several models have been developed in order to retrieve soil moisture using RADAR data.
However, these models need precise prior knowledge about surface roughness. Within this framework, the present
research aims to investigate the capabilities of multi polarimetric RADAR images to overcome the use of in situ data for
surface roughness assessment. The research is carried out on a 24 km² test-site of DEMMIN (Görmin farm),
Mecklenburg Vorpommern, in the North-East of Germany approximately 150 km north from Berlin. Data were acquired
within ESA-funded project AgriSAR 2006 between April and July 2006. Images used include L-band in HH, VV and
HV polarizations acquired from the airborne sensor E-SAR system operated by the German Aerospace Center
(Deutsches Zentrum für Luft- und Raumfahrt - DLR). Two models have been coupled in order to obtain a rms Surface
Roughness Index (rSRI) that is related to terrain physical characteristics as well as vegetation surface properties. These
are the PSEM (Polarimetric Semi-Empirical Model) published by Oh et al. in 2002 and a semi empirical model
developed by Dubois in 1995. A finite difference iterative solution allowed rSRI retrieval without the use of in situ data.
Results have been compared both with in situ rms roughness over bare soil and with Normalized Difference Vegetation
Index (NDVI) obtained from Airborne Hyperspectral Scanner (AHS) optical images collected over the whole
phenological cycle. They show a good agreement with bare soil in situ data, describing its whole range of variability
well, and moreover the NDVI vs. rSRI relationship seems similar to that occurring between NDVI and Leaf Area Index
(LAI) for most crop types meaning that rSRI can be considered as LAI look like.
|
