Unfolding the Luce Deposit to Improve Variograms

Improving the quality of the variogram helps to improve subsequent estimation and simulation processes. Two strategies for conditioning data prior to construction of the variogram are investigated using data from the Luce iron ore deposit in Canada. One strategy is the classic approach in which the deposit is subdivided into a set of structural domains. The domains are analysed individually and datapairs are recombined to produce a variogram. The second strategy is reconstruction of the deposit at an earlier stage of its history through unfolding. This involves transformation of the Cartesian coordinates into stratigraphic coordinate system. Improvement in the variogram quality, expressed in a decreased weighted deviation from a fitted model, is observed for both types of data conditioning. This indicates that, for the Luce deposit, application of data conditioning could also help to inform the selection of sampling locations.