In the last twenty years, important improvements have been shown in backscattering simulations over rough bare soils. However, we still observe for many cases large discrepancies between models and real radar data. In this paper, we propose to illustrate two solutions to improve analysis of radar signal behaviour, by introduction of new approaches for roughness and dielectric constant descriptions. In recent years, the presence of a new type of agricultural surface tillage, used for the sowing of wheat and corn, has been observed with increasing frequency. It illustrates less roughly ploughed soils, with a greater quantity of small clods distributed over the soil surface. In this paper, a new description of such rough agricultural surfaces is proposed. It is based on a composite model, including a classical surface, represented by an exponential correlation function, together with random cloddy structure. This description enables volumetric structures to be introduced over the soil's surface. A numerical moment modelling method, based on integral equations, is used to evaluate the contribution of clods to the radar backscattering behaviour of agricultural surfaces. It is found that the presence of clods explains the very small correlation lengths which are often found in cloddy agricultural fields. The classical approach, in which the surface is described by a correlation function only, based on two statistical parameters, rms height and correlation length, over-estimates the backscattering coefficients when compared to an approach that includes the clods. This overestimation is often observed with real radar data for such fields. For dielectric constant modelling, our analysis is based on a large number of radar measurements acquired during different experimental campaigns (Orgeval'94, Pays de Caux'98, 99). We propose a dielectric constant model, based on the combination of contributions from both soil and air fractions. This modelling clearly reveals the joint influence of the air and soil phases, in backscattering measurements over rough surfaces with large clods. A relationship is established between the soil fraction and soil roughness, using the Integral Equation Model (IEM), fitted to real radar data. Finally, the influence of the air fraction on the linear relationship between moisture and the backscattered radar signal is discussed.