Soil moisture estimation in vineyards by means of C-band radar measurements

Abstract:
A two-component polarimetric model for soil moisture estimation on vineyards developed for C-band radar data is proposed. The key feature of such a scenario is that it is sparsely covered by vegetation and, consequently, it can be modelled by one dominant direct return from the soil and a second multiple scattering component accounting for disturbing and non-modelled signal fluctuations from soil and short vegetation. The polarised direct response from soil is characterised by means of a combined X-Bragg/Fresnel approach, whereas two different methods consisting on mathematical and physical constraints have been considered for the volume component. The model gives a consistent interpretation of the data as long as the volume component power remains about or below 50% of the surface contribution power. Additionally, other two existing strategies originally developed for L-band and further modified for X-band measurements have been also assessed. The key points of such approaches are the physically constrained estimation of the volume intensity by realistic soil characteristics predicted by the Bragg or the IEM (high frequency assumption) scattering model. RADARSAT-2 polarimetric data were used for the assessment of these models concerning soil moisture estimation in vineyards.