SUMMARY
Erosion is dependent on factors such as volume and intensity of rainfall, the infiltration rate of the soil water storage capacity and surface soil water. In this sense, the study of surface roughness of the soil favors the understanding of a number of factors related to the process of soil erosion, favoring thus reducing the damage caused by degradation by soil erosion. On the other hand, you also need to understand how soil surface roughness manifests itself after the incidence of distinct rainfall because the rainfall intensity also plays an important role on the process of soil degradation. The use of geostatistics to the spatial modeling is based in most cases on the assumption of stationarity of data. The objective of this study was to evaluate the spatial distribution of soil roughness in experimental microplots and validate the intrinsic hypothesis of geostatistics. The soil aggregates of 3-5 cm were collected in the city of Lugo (Spain). In the laboratory experimental plots (200 mm x 200 mm, 400 mm x 400 mm, 600 mm and 600 mm x 800 mm x 800 mm) were constructed using metal base mobile 0.86 mx 0.86 m (0.73 m2 ), and the roughness of the ground (dm) measured with rugosimeter laser. Os different size plot for determining the soil surface roughness demonstrated stability in statistical and geostatistical parameters, allowing determination of random roughness index (RR) reliably , regardless of portion size. The validation of the intrinsic hypothesis of geostatistics using data from soil roughness was only possible because the data are homogeneous and represent the same phenomenon in a continuous manner.