Soil Moisture Estimation under Vegetation applying Polarimetric Decomposition Techniques
Thomas Jagdhuber(1), Helmut Schoen(1), Irena Hajnsek(1) and Konstantinos Papathanassiou(1)
(1) German Aerospace Centre (DLR), PO BOX 1116, 82234 Wessling, Germany
In this paper, the potential of using polarimetric SAR acquisitions for the estimation of volumetric soil moisture under agricultural vegetation is investigated. Soil moisture estimation by means of SAR is a topic that is intensively investigated but not yet solved satisfactorily. The key problem is the presence of vegetation cover which biases soil moisture estimates. We discuss the problem of soil moisture estimation in the presence of agricultural vegetation using C- and L-band polarimetric SAR images.
SAR polarimetry allows the decomposition of the scattering signature into canonical scattering components and their quantification. Simple canonical models for surface, dihedral and vegetation scattering using polarimetric decomposition techniques for the interpretation of scattering processes are discussed. Further, the impact of the volume layer modeling by considering different (particle) shape and orientation distributions is emphasized and the role of soil roughness on the dihedral scattering component is investigated. The performance and modifications of the individual scattering components are discussed.
Inversion algorithms were developed to extract soil moisture from the ground components and applied on the data set of the OPAQUE campaign, which was conducted in May 2007 by the University of Potsdam, the German Research Centre for Geoscience (GFZ) and the German Aerospace Centre (DLR) in the eastern part of Germany, south of Dresden. The OPAQUE campaign studies operational discharge and flooding predictions in head catchments and aims to reduce the uncertainties in flood forecasting and prediction of rainfall-runoff processes by identifying critical catchment states caused by saturated soil layers. In the frame of this campaign full-polarimetric and interferometric SAR data (C- and L-band) were acquired by the E-SAR sensor of DLR. Simultaneously, soil moisture was measured with time domain reflectrometry (TDR) probes by the University of Potsdam and GFZ on selected test fields with different vegetation and soil types. In addition, vegetation parameters were collected to characterize the biomass layer.
Potentials and limitations of the algorithms are investigated and soil moisture derived from the inversion of the SAR data is compared to ground measurements for a first quality assessment