Kicking the Bucket - A Case Study of Evolving Domains and Geological Modelling, Sunrise Dam Gold Mine, Western Australia

The Mako domain comprises the Mako lodes at Sunrise Dam Gold Mine, Western Australia. This domain was initially bulk or æbucketÆ domained for the open pit resource in 2000, due to complexity in defining the lodes as a result of low drill density. Negative ore tonnage and positive grade and ounce reconciliations were experienced during mining of the domain in the mega-pit, though only small volumes of the domain were mined. Significant amounts of exploration and underground grade control drilling has been completed since the initial bulk domain interpretation and greater geological understanding of the mineralisation controls and geology of the lodes now exists. Mako provides approximately 50 per cent of the ore tonnes and ounces for the North Wall Cut-back, the latest open pit phase of mining at Sunrise Dam. The importance of Mako to mill supply and life-of-mine planning, as well as past reconciliation performance necessitated a reinterpretation of the domain to increase confidence for mining. Historic exploration drilling, underground grade control drilling and detailed pit-wall mapping were used to build mineralised subdomains within the confines of the original domain. These reflect more closely the spatial extents and geological controls on the mineralisation. A new resource model using the multiple indicator kriging (MIK) technique was generated using the new domains. It resulted in a decrease in tonnes and ounces but increase in grade for the sum total of the new domains compared to the old æbucketÆ domain. This is attributed to the increased selectivity and structural constraints imposed by the new domains. A retrospective comparison of the Mako grade control from the mega-pit versus the new resource model was conducted, which æreconciledÆ 103 per cent tonnes, 106 per cent grade and 110 per cent ounces. This has provided confidence that the new model is a better estimate of the lodes. Qualitatively, the new model is a better representation of the geology and extent of the mineralised structures with significant areas of change identifiable. Grade-tonnage plots comparing the 2006 and 2008 models show that most tonnes have been shed from the low-grade û subeconomic grade cut-offs. This is mostly attributable to the volume reduction of æoreÆ in the 2008 model compared to the 2006 model, where the true mineralised zones are captured by the new domains. The reinterpretation and remodelling of Mako is an example of an evolved domain model and shows that mineralised zones need to be differentiated adhering to the geology and geological controls on mineralisation, which are understood largely due to increased drilling density coupled with mapping.