Changes in dispersion variance consequent upon inaccurately modelled semi-variograms

The semi-variogram function which describes the spatial variation of samples as a function of their distance apart is an intermediary in geostatistical calculations. It is used in the calculation of several variances: the estimation variance which quantifies the accuracy of block estimates and the dispersion variance which measures the variability of blocks in the deposit. However the semi-variogram must be accurately modelled if these variances are to be well-known. The changes in dispersion variance calculated here when a spherical semi-variogram is estimated inaccurately show the need for care in its estimation, especially in the selection of the correct nugget effect.