Geology-aided Geostatistical Modelling of a Ferruginous Bauxite Deposit in Eastern India

Abstract

Geostatistical modelling for a part of a ferruginous bauxite deposit in Eastern India has been carried out for estimation of mineral inventory and grade-tonnage relationships. Resource estimation of the bauxite deposit carried out by different agencies previously were on the basis of drill hole spacing employing fundamentally conventional cross-sectional techniques and in a limited way geostatistics. The conventional methods have limitation to specify the confidence limits (+/-) or, in other words, the percent accuracy of the estimated mean. One normally relies on approximation without any mathematical basis. However, no integrated mineral deposit modelling has been carried out combining deposit geology with geostatistics for development of mineral inventory and estimation of grade-tonnage relationships. An understanding of formation of bauxite by the process of residual weathering and lateritization of Archean khondalite and charnokite, when incorporated in population modelling and semi-variography and integrated with appropriate geostatistical evaluation process provided a means to improved geo-characterization of the bauxite deposit. The integrated geostatistical modelling led to the quantification of the individual population characteristics and spatial relationships of the geo-variables. Spatial correlation modelling of the sample values through semi-variography revealed a moderate to high nugget-to-sill ratio with moderate range of influence, characterizing the spatial variability of the ferruginous bauxite deposit. Geostatistical estimation employing 3D block kriging led to the generation of slice-wise block kriged estimates and kriging variances. Resulting block estimates, when stacked slice-wise one below the other, provided a 3D mineral inventory of 29.87 mt of bauxite averaging 42.04% Al2O3 and 2.88% SiO2. Spatial distribution maps of kriging variance aid in reflecting zones of uncertainty of varying magnitude associated with block estimates.