Geostatistical Modelling Of An Australian Iron Ore Body

One of the more difficult problems frequently facing geologists and mining engineers is that of estimating the ore reserves of mineral deposits. Cost conscious mining operations invariably require an accurate knowledge of tonnage and grade for both short and long term planning. A case study is presented, showing how a classical frequency distribution was fitted to a population of drillhole assays. Using these results, a grade- tonnage curve is derived using the Lasky expressions, and the results discussed. A geostatistical analysis is then presented. First, variograms are drawn, models fitted to them and analyses made of drift and anisotropy. Second, simple block kriging is carried out and the results plotted on large 1000 scale plans, suitable for overlay with routine .mine plans. Kriging results are verified using two independent techniques: random kriging on a regular grid and inverse distance weighting. Finally an attempt is made to optimize the drilling grid, and to fully tabulate the total computer running costs for the work described.