Papers

Summary

Regional studies where geostatistic is applied for describing landscape models have shown that, even in places previously considered homogenous, different spatial dependencies have been found and these had been related to crop productivity. In order to analyze the physical properties spatial variability models of an eutrudox Red Latosol, a regular grid containing 306 points spaced at separation distances of 50 meters was installed in southern Brazil. Soil samples were collected at 0-0.2 m and 0.6-0.8 m in order to determine clay, silt, sand total, sand thick, sand medium, sand fine and sand very fine contents. Data were submitted to descriptive statistics, geostatistic and in sequence to kriging analyzes. The coefficient of variation of studied properties was low for clay and, medium for silt, sand total, sand medium, sand fine and sand very fine and, high for sand thick. All the spatial variability models had moderate variability dependence, with higher ranges for 0-0.2 m of depth. The results showed that the considered homogeneous characteristics of oxisols submitted to the same soil management is not so evident and spatial variability of texture attributes can be distinguished from each other.