Estimated values of soil particle size distributions using cokriging

Abstract

Using the multivariate technique of kriging known as cokriging, it possible to estimate values of variables high cost of aquisition using variable of the low cost and fast acquisition, provided that both exhibit high spatial correlation. The objective of this study was to evaluate the use of electrical conductivity of 1:5 extract (ECe 1: 5) as a covariate in an analysis by cokriging to estimate the values of granulometric components of an Oxisol. In an area of 90 ha, soil samples were collected at 181 georeferenced points, at depth of 0 to 0.20 m. The samples were sent to the laboratory for characterization of soil particle size (coarse sand, fine sand, silt and clay) and measurement of ECe (1:5). Using the cokriging, we attempted to estimate the values of the variables soil texture using the ECe (1:5) as a covariate. Using the technique of cokriging, the best estimates of values for unsampled locations were obtained for the variables silt and clay, because they showed greater spatial correlation with the electrical conductivity of saturation extract (1:5).