The various spatial interpolation techniques used offer different performances, according to the characteristics of the initial data and it is very common to find evaluators relying only on R² (coefficient of multiple determination) and residual error. This study aimed to apply the Taylor diagram to analyze 15 different methods of spatial interpolation in area with commercial cultivation of eucalyptus, to compare and choose the best method of interpolation of a set of values of the size granulometric fraction of the soil (clay) at the depth of 0-0.20 m from the soil. According to the results, the diagram Taylor tool was considered very satisfactory, because in addition to graphically define the best interpolation methods, it allows to make choices among them within a smaller and more defined set of information; therefore, the Splines method was chosen rather than linear kriging model. The statistics of these two methods were very similar, with minor variations, and the standard deviation of the Splines was closer to the observed data, which means that it is the best interpolation method for clay at a soil depth of 0-0.20 m.
Graphics interpolators; Spatial distribution; Soil texture
