Surface roughness is an important geo-physical parameter required for numerous applications such
as agronomy, geology, risk assessment, etc. In addition, the estimate of soil roughness may provide
valuable a priori information to simplify the problem of soil moisture retrieval from SAR data.
In the past, roughness discriminators based on the ratio between soil backscatter at different
) and on the correlation coefficient between
HH and VV channels (i.e.
have been suggested. More recently, the potential of the correlation coefficient between
co-polarised channels (i.e. polarisation coherence) in an arbitrary state of polarisation has been
investigated. In particular, the correlation coefficient between co-polarised channels at circular
has been found extremely sensitive to surface roughness and weakly
sensitive to soil moisture content. However, notwithstanding these observations have been
confirmed by several experimental studies a complete physical understanding of the phenomenon is
still missing, at least in the remote sensing community.
One of the main reason for this lack of understanding is that in general, only lowest order
approximations of theoretical surface scattering models are exploited in remote sensing
applications. These approximations do not include the effect of multiple reflections. They cannot
therefore predict accurately the whole covariance matrix often required to synthesise roughness
discriminators, such as
rhoRRLL. In this respect, despite the fact that higher order approximations of
theoretical surface scattering models are mathematically very complex, they are necessary to give
indications to understand the phenomenon and they can provide physical guidelines to develop
In this context, the objective of this paper is to present a
simple physical framework to interpret the
sensitivity of different roughness discriminators to soil roughness. The adopted interpretation
scheme is based on indications provided by 2nd order approximations of surface scattering models,
such as Small Perturbation Method (SPM), Small Slope Approximation (SSA) and Kirchhoff
Finally, limits and perspectives of using ASAR and the forthcoming PALSAR data to retrieve
surface roughness is discussed.