Soil moisture mapping in a semiarid region, based on ASAR/Wide Swath satellite data

In this paper, an operational algorithm is proposed for the mapping of surface moisture over the northern and central parts of Tunisia, in North Africa. A change detection approach is applied, using 160 multiincidence Envisat ASAR Wide Swath images acquired in the horizontal polarization over a 7 year period. Parameterization of this algorithm is considered for three classes of vegetation cover density (NDVI < 0.25, 0.25 < NDVI < 0.5, and NDVI > 0.5), retrieved from SPOT-VGT decadal images. A relative soil moisture index, ranging between 0 (for the driest surfaces) and 1 (for saturated soils), is proposed for each date, with a resolution of 1 km. The retrieved soil moistures are validated by means of ground measurements based on continuous thetaprobe measurements, as well as low-resolution (25 km) ERS and ASCAT soil moisture products from the Vienna University of Technology (TU Wien). A qualitative relationship between spatiotemporal variations of moisture and precipitation is also discussed.