Using Geostatistics to Assist in Optimising Grade Control Estimation at KCGM's Fimiston Open Pit

The Golden Mile, which has been mined continuously for the past 100 years by both underground and more recently open pit methods, occupies an area of 1.2 km by 4 km adjacent to the city of Kalgoorlie Boulder in Western Australia (Figure 1). Large-scale open pit mining at the Kalgoorlie Consolidated Gold Mines (KCGM) Fimiston operations is currently extracting 7.8 million tonnes of oxide and sulphide ore for a total production of 522 000 oz per annum. The effectiveness of grade estimation techniques at KCGM has been examined with reference to existing 10 m bench sample data from all operating areas within the Eastern Lode System of the Main and North pits. The aim of the study was to recommend practical parameters for grade control estimation. The review included comprehensive variography and comparison of nonparametric estimation methods (Joumel, 1983) such as indicator kriging and probability kriging with ordinary kriging, leading to the development of improved grade control parameters and procedures for both daily quality control and short term mine planning. This paper reviews the geological, statistical, geostatistical, mining and modelling features studied in order to present KCGM's Mining Department with recommendations for improving grade control.