A variogram is a geostatistical tool that quantifies and models the spatial continuity, correlation, and variability of a geological attribute — such as gold grade, iron content, aluminum oxide, or diamond carat frequency — as a function of the distance and direction between sample locations within a mineral deposit. It is a fundamental input to kriging and other geostatistical estimation techniques used in the construction of mineral resource models for bauxite, gold, iron ore, and diamond deposits. The variogram is computed by calculating the average squared difference in attribute values between pairs of samples at various separation distances (lags) and plotting these values against the lag distance to produce the experimental variogram. A mathematical model — commonly a spherical, exponential, or Gaussian function — is then fitted to the experimental variogram, defining parameters including the nugget effect, sill, and range. The nugget effect represents the variability at zero distance, attributable to micro-scale variability and sampling error; the sill represents the total variance of the dataset; and the range represents the distance beyond which there is no spatial correlation between samples. In gold mining, variogram analysis is particularly important because of the high variability typical of gold distributions, where the nugget effect can dominate the variogram and significantly influence estimation uncertainty. In iron ore and bauxite deposits, which tend to exhibit smoother grade distributions, variograms often display well-defined ranges and structures that facilitate reliable grade interpolation. In diamond resource estimation, the rarity and discrete nature of diamond occurrences poses specific challenges for variogram computation and modeling. The variogram underpins all kriging estimates and directly influences the quality and reliability of mineral resource models.
