The Use of Cokriging with Limited Field Soil Observations

Abstract Spatial variability of sand, silt, and clay contents, available water content (AWC), and water stored at ⅓ bar (pF2.5) measurements was studied by using classical statistical and geostatistical concepts. The measurements were made within a 70‐m long and 40‐m wide area at the nodes of a 10‐m square grid. Semivariograms for all the variables and cross semivariograms for the spatial correlation between AWC, pF2.5, and sand content values were used to krige and cokrige additional values of AWC and pF2.5 every 5 m. It is shown that the use of cokriging is a promising tool in soil physics, whether the principal objective is the reduction of the estimation variance compared with kriging, or the need to estimate one under‐sampled variable by taking into account its spatial correlation with another better‐sampled variable. Finally, the results of kriging and cokriging have been compared with the classical statistical treatment in which the variables of interest (AWC, pF2.5) at unobserved points were estimated by simple linear correlations with another variable (sand content) measured at those points.