Confronto critico tra kriging e analisi oggettiva

Equicontour-line maps are often used in the reconstruction and in the graphic display of environmental phenomena. Effective and reliable interpolator tools are required to do this in the best objective manner. The stochastic-interpolation procedure based on the kriging technique satisfied this requirement and it is mostly applied to geological and hydrological problems. Along with the equicontour-plot map, the map of the standard deviation σ is given and provides the degree of reliability of the variable field reconstructed. A similar interpolation procedure, commonly used in meteorological-oceanographical fields, is called objective analysis. Both methods objectively reproduce the event under study, based upon two principal statistical hypotheses: the covariance of experimental data (variogram) and the trend of the occurrence. In this paper the two techniques are compared, either in their theoretical development, or in their application to the two different experimental data (geological and oceanographic). From the discussion which follows it can be concluded that the two methods are substantially identical. The objective analysis shows however more generality because it could be applied to time-dependent fields. © 1993 Società Italiana di Fisica.