Resumo

ACEVEDO, D. Cristóbal et al. Spatial Variability of Chemical Properties of the Soil and its Use in the Design of Experiments. Terra Latinoam [online]. 2008, vol.26, n.4, pp.317-324. ISSN 2395-8030.

Spatial variability of the chemical properties in the soils is a problem which the researchers face when working in precision agriculture or when they want to know how nutrient concentrations, or some other useful chemical parameters, are distributed for selecting experimental sites. In the present research the spatial variability of the chemical parameters of the soil was characterized [pH, organic matter (OM), total nitrogen (Nt), available phosphorus (P-Olsen) and exchangeable potassium (K)], using classic statistics (variance and coefficient of variation) and geostatistics (semivariance), to obtain some representative values and maps that show their behavior. Soil sampling (0-20 cm depth) was done in a grid with equidistance of 25 m, obtaining 182 samples in an agricultural table with an area of 12.7 ha. With the data obtained the spatial variability of the soil was analyzed, obtaining the following average values, coefficient of variation (CV), and minimum number of determinations to obtain its average value with 95% of confidence (n): 6.97, 3.7%, and 2 for pH; 2.3%, 25.5% and 103 for OM; 0.115%, 24.3% and 93 for Nt; 21.34 mg kg^-1, 53.6% and 451 for P-Olsen; 215.97 mg kg^-1, 70% and 768 for K. When applying the geostatistics approach, all the parameters presented spatial structure, declared in the experimental semivariogram, adjusting the variables to the theoretical spherical Gaussian models, linear, linear and hole effect, for pH, OM, Nt, P, and K, respectively. Distribution maps of the parameters were obtained applying the kriging method and zones with different gradients and important patterns of variability for the establishment 317of experiments and the practice of precision agriculture were located.

Palavras-chave : semivariogram; clasic statistics; geostatistics; sampling; kriging.