Two-Dimensional Geostatistical Methods ùTheory, Practice and a Case Study from the 1A Shoot Nickel Deposit, Leinster, Western Australia

Two-dimensional (2D) estimation approaches have significant advantages over three-dimensional (3D) approaches when dealing with a narrow vein- or layer-like geometry. The main motivation for 2D estimation is that bias due to non-additivity of grade is avoided. The various aspects of 2D estimation are illustrated by use of a case study from the 1A Shoot Nickel deposit at Leinster, Western Australia. The width of the 1A mineralisation is variable, and so are the angles of drilling intersection. Consequently, the variable of interest (in this case nickel grade) is not suitable for direct kriging because it is defined on non-uniform support. However, grade can also be defined as the ratio of two variables that are suitable for direct kriging: thickness and accumulation. In the 2D approach, tonnage equates to thickness and metal equates to the product of grade and thickness, also called æaccumulationÆ. In the simplest approach we can independently krige accumulation and thickness and derive the grade by the ratio of the two resultant estimates. The thickness and accumulation variables must be in a consistent projection orientation, for example a horizontal projection in the case of steeply inclined mineralisation like 1A. Because the density is not constant across the 1A (and most other) deposits, the issue of density weighting is also discussed. The impact of the correlation between grade and length, or grade and density are important issues in implementation of 2D estimation. The circumstances under which these correlations need to be taken into account (by co-kriging) are reviewed. Finally some comments are made on the final use of the resultant 2D estimate for mine planning purposes.