Internal Consistency and Inference of Change-of-support Isofactorial Models.

Bivariate isofactorial models are used for global or local change-of-support applications. However, so far, their variogram analysis is complicated and may lead to mathematical inconsistencies. In this paper, we propose an alternative approach for internally consistent variogram inference, which consists in deriving the simple and cross variograms at point and block supports from the variogram of transformed data, by randomizing the sample locations within the blocks (regularization). This approach is illustrated with the discrete Hermitian model, for which we provide guidelines for parameter inference and emphasize the limitations of the extreme cases: discrete Gaussian and mosaic models. A case study is presented with an application of the Hermitian model to a mining dataset, which consists of drillhole samples measuring the grade in a porphyry copper deposit. [ABSTRACT FROM AUTHOR]