In precision agriculture management principles must be adapted to the field variability. This requires efficient techniques to estimate and map the spatial and/or temporary variability of soil attributes and properties. However, the determination of some variables that characterize the properties of a particular soil is often onerous and troublesome. In these situations it is interesting to estimate such variables as a function of others that present good space correlation with the forner and are of simpler determination. This is possible by a cross-semivariogram. The interpolator that uses the cross-semivariogram in modeling is called co-kriging. The aim of this study was to test the co-kriging method for estimating pH and Mn as related to soil organic matter to obtain the error range associated to this technique and to compare the estimated values with those determined in laboratory. Results showed that soil variability can be estimated with high precision by co-kriging.
geostatistics; spatial variability; precision agriculture
