EVALUATION OF POLARIMETRIC DECOMPOSITION FOR SOIL MOISTURE RETRIEVAL OVER VEGETATED AGRICULTURAL FIELDS Hongquan Wang 1 , Ramata Magagi 1 , Kalifa Goita 1 , Thomas Jagdhuber 2 and Najib Djamai 1 1 CARTEL, University of Sherbrooke, 2500 boul. de l'Université, Sherbrooke, QC J1K2R1, Canada 2 German Aerospace Center, PO BOX 1116, 82234 Wessling, Germany ABSTRACT This study presents a simplified polarimetric decomposition for soil moisture retrieval under vegetation cover. After removing the volume scattering contribution in the full polarimetric SAR signature, only the surface scattering component is used to retrieve the soil moisture. The simplified algorithm is evaluated on the dense time series of UAVSAR data and detailed ground truth measurements covering the whole crop growth period. The results show that the performance of the soil moisture retrieval depends on both the crop types and the phonological developing stage. The fields covered by soybean obtain better results than other crop type, due to the low crop height and biomass. This study validate the potential of polarimetric decomposition for soil moisture retrieval over vegetated agricultural fields. Index Terms— Polarimetric decomposition, Soil moisture, SMAPVEX12, UAVSAR 1. INTRODUCTION Soil moisture is an important factor in several physical process (such as water conservation, soil erosion and surface runoff) over agricultural fields. Compared with the conventional soil moisture acquisitions by point sampling which is time and labour consuming, Synthetic Aperture Radar (SAR) has the potentials to extract subsurface information with high spatial and temporal resolution. Furthermore, polarimetric SAR extends the observation space, providing the potentials to improve the robustness of soil moisture retrieval. Nevertheless, agricultural fields are covered by different crops over large periods annually. In this condition, the effects of vegetation layers and the underlying soils are coherent superimposed in the measured SAR signature, which complicates the problem of soil moisture retrieval over vegetated soils. In order to solve this problem of vegetated soil moisture retrievals from SAR measurements, intensive studies are conducted in the past decades [1-3]. On one hand, the vegetation layer is accounted and removed by incoherent radiative transfer model [3]. On the other hand, based on the polarimetric SAR measurement, the target decomposition methodology [4] is adopted to decompose the full signature into different scattering mechanisms (e.g. single scattering, dihedral scattering and volume scattering). Thus, the vegetation component is removed by such polarimetric decomposition and then the underlying soil moisture could be derived from the ground scattering component. Within the above context, this study is dedicated to solve the problem of soil moisture retrieval under vegetation cover, based on a simplification of the polarimetric decomposition algorithm in [1]. Under the framework of SMAPVEX12 [5], the dense time series of UAVSAR data were acquired with the detailed ground truth measurement of soil and vegetation characteristics. As the encouraging results are obtained in [1, 2] over vegetated agricultural field by model based polarimetric decomposition [1], the objective of this study is to evaluate whether this polarimetric decomposition approach is applicable on the SMAPVEX12 experimental datasets. In this paper, the section 2 describes the used datasets. The method of simplified soil moisture retrieval is proposed in section 3. The main results are analyzed and discussed in section 4 and the main conclusion is presented in section 5. 2. DATASET PRESENTATION The SMAPVEX12 study site [5] (with dimension 15 km × 70 km) is located within the larger Red River Watershed in Canada. This site is dominated by several annual crops, including cereals (32.2% of area), canola (13.2%), corn (7.0%) and soybean (6.7%). In addition, 16.4% of the site is occupied by perennial cover (grassland and pasture), and some area is also covered by forest. Furthermore, the soil texture varies across the site, and this spatial soil variability provides a large range of soil moisture conditions. The diversities of land cover (especially different crop types) and soil types facilitate the development of the SMAP soil moisture retrieval algorithms under vegetation cover condition. Over the study site, UAVSAR data acquisition covered 14 dates from June 17 th to July 17 th 2012, with the following characteristics: L-band, incidence angles ranging from 25° to 689 978-1-4799-7929-5/15/$31.00 ©2015 IEEE IGARSS 2015