Kriging is a sophisticated geostatistical interpolation technique widely used in the mining industry — across gold, iron ore, bauxite, and diamond deposits — to estimate the grade, thickness, or other geological attributes of a mineral deposit at unsampled locations, based on the values measured at sampled data points (drill holes, channel samples, or grab samples) and the spatial correlation structure of the variable of interest, quantified through a mathematical model called the variogram or semivariogram.
Named after the South African mining engineer Danie G. Krige, who pioneered spatial estimation methods in the Witwatersrand gold mines of South Africa in the 1950s, kriging provides the Best Linear Unbiased Estimator (BLUE) of a spatially distributed variable — meaning it minimises the estimation variance (error) while ensuring that estimates are unbiased (neither systematically high nor low). The fundamental tool of kriging is the variogram, which describes how spatial correlation between sample pairs decreases with increasing separation distance, characterised by three key parameters: the nugget (random variability at zero distance), the sill (total variability), and the range (distance at which spatial correlation effectively ceases). Several variants of kriging are employed in resource estimation practice, including Ordinary Kriging (OK), which is the most commonly used, Simple Kriging (SK), Indicator Kriging (IK) for estimating grade-tonnage curves, and Kriging with External Drift (KED) for incorporating secondary information. In gold deposits with high nugget effect, kriging must be applied carefully to avoid conditional bias, and grade capping or top-cutting of extreme assay values is often required prior to estimation to ensure robust resource classification under the JORC Code or NI 43-101 reporting standards.