In this paper a novel multi-scale roughness description is presented and it
s impact on surface backscattering is investigated. The analysed rough surf
aces are synthesised by exploiting a Karhunen-Loeve decomposition of 1/f ra
ndom processes in terms of wavelet bases. For these surfaces, relevant roug
hness parameters are not traditional quantities like the height profile std
(s) and the profile correlation length (I) but they are a new set of parame
ters related to the surface multi-scale properties. Besides, the paper inve
stigates the possibility of matching the multi-scale roughness description
with e.m. asymptotic models such as the Integral Equation Method (IEM). Und
er appropriate hypothesis, it is shown that the IEM model can be adapted to
the multi-scale roughness description and a sensitive analysis of copolari
sed backscattering coefficients to relevant roughness parameters is carried
out in the case of one dimensional surfaces.