Suitability of Different Observables and Polarizations for Dinsar-based Soil Moisture Estimation

The mapping of soil moisture at field scales can potentially be achieved by the remote sensing technique differential interferometry, or DInSAR. Its signals have been shown to be sensitive to soil moisture changes. We analyse the suitability for this task of three DInSAR observables: the phase φ, the coherence |γ| and the phase triplets Ξ. By inverting an electromagnetic scattering model, we obtain median correlations between the estimated and the in-situ measured soil moisture time series of 0.86 for the HH phase φHH , 0.85 for the coherence |γHH |, and 0.78 for the phase triplets ΞHH . The phase observable is also sensitive to deformations: these predominantly impact the phase in a way similar to soil moisture changes, thus rendering the separation difficult, e.g. for providing corrections for deformation studies. The other two observables are not affected by deformations, but turn out to be less robust and insensitive to certain temporal soil moisture variations. In the absence of deformations the phase is thus the most promising DInSAR observables for soil moisture retrieval.