Effect of Geological Models on ML/ARD Characterization Program Design at the KSM Project

摘要

The effect of different geological classification systems on a ML/ARD characterization program is investigated for the KSM Au-Cu porphyry project in northwest BC Canada. Geological models enable predictive modeling of the ML/ARD potential of material during excavation and long term storage of waste rock by allowing the representation of a large mass of material by a relatively small mass of sampled material. A robust sampling plan typically attempts to collect samples in proportion to the mass of material to be excavated. Geological models developed during advanced exploration are often used during the geochemical characterization program; however, biases towards ore at the expense of waste rock and future pit walls frequently exist. Perceived uncertainty introduced through "lumping" or "splitting" of datasets can be carried into ML/ARD prediction and water quality models and regulatory approvals. Therefore, the choice of geological model is essential as a geochemistry baseline program advances to determine quantities of potentially acid generating (PAG) and not-PAG waste rock for waste rock and water management planning and to address stakeholder and government concerns.Three geological models of the KSM project were assessed: Lithology, Alteration, and a hybrid Mine Model. Using the proportion of samples classified as PAG in each model unit, the material to be excavated was classified as PAG or not-PAG.These results were then compared to an ABA block based on each of the three geological models. Differences in the proportions of PAG material were identified and the underlying cause of the differences is presented.