The Use of C- and L-Band Repeat-Pass Interferometric SAR Coherence for Soil Moisture Change Detection in Vegetated Areas

ABSTRACT

Soil moisture estimation studies using Synthetic Aperture Radar (SAR) routinely utilise only the amplitude part of the received echo. In this study, repeat-pass C- and L-band interferometric SAR coherence from 2007 - 2009 was evaluated for the detection of surface soil moisture changes in the presence of vegetation using two different approaches. In the first analysis, the association between low coherence and large in situ soil moisture changes was investigated using 24 interferometric pairs and the decorrelation effects due to vegetation and weather were also assessed. Results reveal that, in very few cases soil moisture differences between acquisitions contributed to the signal decorrelation. For the majority of cases, particularly in C-band, the change in vegetation tended to be the predominant source of decorrelation, suppressing the influence of any soil moisture changes. The second analysis applied thresholds to both coherence and intensity data to determine if a coalesced coherence (γ) and intensity change (Δσ⁰) approach could improve detection of changes in measured soil moisture content. The aim was to test the usefulness of a γ > 0.3 and Δσ⁰ > 1.5 dB thresholding approach to separate the effects of a vegetation change and a soil moisture change on the SAR signal. Results suggest that the approach improves the reliability of the soil moisture change detection although clearly limits the use of available image pairs. These analyses demonstrate the increased information the coherence adds to SAR studies over agricultural areas.

Keywords: