Integrating Multivariate Geostatistical Simulation and Lithology-Based Geometallurgical Parameters for Ultimate Pit Optimization

This study presents an integrated methodology to incorporate geological uncertainty and geometallurgical variability into ultimate pit optimization for an iron ore deposit in the Quadrilátero Ferrífero, Brazil. The proposed framework integrates multivariate geostatistical simulation, lithological grouping, and economic optimization. The deposit was classified into three principal lithological groups (friable, semi-compact, and compact), while less representative lithologies were assigned to a residual “Others” category. Iron grades (Fe1–Fe4) and particle-size fractions (g1–g4) were simulated using the Turning Bands Simulation, generating 25 equiprobable realizations. Lithological grouping defined mining and processing costs, as well as mass and metallurgical recoveries, which were incorporated into Lerchs–Grossmann ultimate pit optimization. The simulated realizations successfully reproduced the statistical distributions, spatial continuity, and multivariate dependence structure observed in the original dataset, enabling the quantification of geological uncertainty and its economic implications. The P50 undiscounted cash flow from realization-specific optimizations was USD 17.52 billion compared with USD 16.36 billion for the deterministic Ordinary Kriging model, a 7.13% difference. This difference does not represent an achievable increase in project value, but reflects the sensitivity of optimized economic outcomes to geological variability and different ultimate pit designs.