24th INTERNATIONAL MINING CONGRESS OF TURKEY, Antalya, Türkiye, 14 - 17 Nisan 2015, ss.250-256
In predicting second-order stationary vector random fields, simple and ordinary cokriging techniques are widely used. In this study, an alternative multivariate kriging technique Covariance Matching Constrained CoKriging (CMCoK) is introduced and compared with ordinary kriging (OK) and its multivariate counterpart ordinary cokriging (CoK). For comparison, Meuse River data are used and results are discussed by means of local accuracy and spatial variability. From the results, it is seen that error variance for CMCoK is considerably less than CoK and OK.