Geometallurgical modelling is now a hot topic in deposit evaluation and in optimising mine performance; however, its history is long and varied._x000D_
Historically metallurgical sampling has frequently suffered from a lack of data due to the cost of producing representative bulk samples. In the early 1990s geometallurgical modelling similarly suffered from a lack of data due to complex, time consuming and expensive sampling techniques. More recently, simple and relatively cheap tests for geometallurgical variables such as hardness, grindability, throughput, SAG power index, bond work index, crushing index, mineral recovery and concentrate grade have become widespread. The cheaper sampling has resulted in relatively large geometallurgical databases. However, most of the geometallurgical variables sampled are implicitly (and incorrectly) assumed to be additive when modelled by traditional geostatistical techniques, often leading to ill-informed and costly decisions. Geometallurgical variables are mostly non-additive so the current geostatistical challenge is to develop new approaches to dealing with this type of data._x000D_
This paper briefly describes a historical case study of geometallurgical modelling of the George Fisher Pb-Zn mine completed in the mid 1990s, then outlines some of the current misconceptions of geometallurgical data and modelling techniques regularly applied to it, and finally discusses some recent developments in geometallurgical modelling, with examples from a number of current clients indicating the direction forward._x000D_
FORMAL CITATION:Richmond, A and Shaw, W J, 2009. Geometallurgical modelling - quo vadis? in Proceedings Seventh International Mining Geology Conference 2009, pp 115-118 (The Australasian Institute of Mining and Metallurgy: Melbourne).