Coherence Characteristics of Radar Signals from Rough Soil
To understand the mechanisms of decorrelation in interferometric SAR (InSAR) images of bare soil, a model has been developed. Under the Kirchhoff and stationary phase approximations, coherence can be related to the statistical variations of dielectric constant and roughness parameters of surfaces. With the help of an empirical model for the dependence of dielectric constant on soil moisture, coherence due to the inhomogeneity of soil moisture is numerically demonstrated. It has been shown that the decorrelation of the radar signal from rough soil is mainly due to the moisture variability within the resolution cell. The effect of roughness on decorrelation is complex. The effect is negligible compared to that of the dielectric variability for homogeneous resolution cells (no dielectric variability within a resolution cell). However, the coherence depends strongly on the roughness parameters for resolution cells with large moisture variability. It is concluded that the loss of coherence induced by variability of dielectric constant can be related to the relative variation of moisture expressed by the ratio of standard deviation and mean value, and that large relative variations of moisture could lead to much decorrelation. If the moisture variability is small the coherence will be very high even if the values of mean moisture of the two SAR observations are different, which means that coherence can be high in spite of much backscatter differences.