博客來Turn drillhole data into defensible precious-metal resource models with modern geostatistics that holds up under scrutiny and reconciles against reality. This engineering-focused reference walks through the full chain of spatial modeling for gold, silver, and platinum group metals, from sampling support and domaining decisions to variography, kriging and co-kriging, conditional simulation, and uncertainty-driven grade-tonnage forecasting.
You will learn how to build variogram models that reflect real deposit geometry and sampling behavior, including high nugget effects, anisotropy, thin tabular reefs, and narrow vein systems. The book then connects those spatial models to practical estimators, showing when ordinary kriging is sufficient, when trends require drift handling, and how multivariate modeling can improve estimates across correlated metals and auxiliary variables without introducing hidden bias. Nonlinear approaches are treated in depth, including indicator methods, conditional distributions, uniform conditioning, and simulation workflows used to quantify risk at SMU, panel, and reporting scales.
A key theme is decision quality under uncertainty. You will see how to move beyond single ＂best estimates＂ into probability-based ore/waste decisions, P10/P50/P90 outcomes, cutoff sensitivity, and uncertainty envelopes on grade-tonnage curves. The final workflow focus is reconciliation, linking model assumptions to mine production and plant accounting, diagnosing systematic sources of bias (support mismatch, domaining, top-cuts, smoothing, dilution and ore loss), and updating models with feedback loops that improve predictability over time.
Every chapter includes complete Python code demos to make the methods reproducible, auditable, and immediately usable in real modeling pipelines, including variogram fitting and validation, kriging and co-kriging, conditional simulation, CCDF estimation, uncertainty metrics, grade-tonnage curve generation, and reconciliation analytics.
