Soil moisture evaluation by means of multi-temporal ERS SAR PRI images and interferometric coherence

The feasibility of using multi-temporal spaceborne Synthetic Aperture Radar (SAR) imagery for the evaluation of soil moisture at a field and regional scale is evaluated through the use of European Remote Sensing Satellite (ERS-1 and ERS-2) images. Six pairs of Single Look Complex (SLC) images and six Precision Images (PRI) were acquired during the 1996 Tandem mission for the test site of Gembloux, in Belgium. The SLC images of each Tandem pair were processed by the "Centre Spatial de Liege" (CSL) in order to produce interferometric coherence images, and the PRI were calibrated; all the images were coregistered and rectified. The backscattering coefficient values and the mean coherence values were calculated for each pilot field and compared with soil moisture and crop characteristics measurements. Our study confirms that for bare soil fields, a Linear relationship exists between the volumetric near-surface soil moisture and the backscattering coefficient for unsaturated soils, with a determination coefficient of 0.75. The innovative key point of the study shows an excellent correlation between the backscatter value of the image subset of the test site (80 sq. km), and the corresponding mean soil moisture measurements, with a determination coefficient of 0.98. The image subset comprises mostly agricultural fields but also forests, meadows and urban areas. Moreover, the relation is as valid in the spring, before the growing season, as in the summer, when the vegetation cover is high. The study also reveals that coherence images are useful to identify bare soil fields. (C) 2000 Elsevier Science B.V. All rights reserved.