A simple kriging method incorporating multiscale measurements in geochemical survey

In this study, we propose a kriging algorithm, multiscale kriging model, to incorporate geochemical data observed at multiscales (multiresolutions). We assume that there are a number of measurements at different scales, and that the target scale at which the parameter values are needed may be different from the measurement scales. Several synthetic examples and the vanadium geochemical data from 8402 stream sediment samples in Zhejiang Province, China, have been used to illustrate the method. These examples demonstrate that, by incorporating measurements from all scales, the estimated field is better than the field estimated using measurements from any individual scale. This method also allows us to estimate a parameter field at the scale that does not have any measurements.