Knowledge about the variability of soil physical properties is fundamental for the monitoring of soil quality for being directly related to the sustainability of agricultural production. This study aimed to: a) assess and identify structures of spatial dependence by the construction of semi-variograms, for a Red-Yellow Latosol (Oxisol) under coffee, b) characterize the spatial distributions of the physical properties using ordinary kriging, and c) identify the layer with greatest physical impediment. The experiment was conducted on an experimental farm of the Agricultural Research Company of Minas Gerais (Epamig), in Patrocínio, MG, on a clayey Oxisol. Samples were collected from trenches at the intersections of a rectangular 40 x 150 m grid, totaling 28 sampling points, where soil samples were collected with metal rings (2.54 cm in height and 6.30 cm in diameter), with an Uhland sampler. Each trench had three steps with a surface of 2.0 x 1.5 m, at the depth of 0.00, 0.10, and 0.25 m, from each of which seven soil samples were collected, arranged in a matrix form, totaling 588 samples. Except for total porosity, soil physical properties were spatially dependent in all layers. The ordinary kriging technique proved to be a viable alternative for the estimation of data points not sampled in the experimental area. The physical restrictions to root growth were most severe in the 0.00-0.03 m layer due to the highest soil density and lowest soil macroporosity.
compaction; geostatistics; soil property maps
