The Geology Twin: Evidence-Bound Dynamic Resource Modelling for Governed Mine-to-Market Decisions

Volume 1 presents the Geology Twin as a governed, evidence-linked framework for geological observations, sampling, laboratory results, QA/QC, structural interpretation, block and geometallurgical models, uncertainty, and controlled geological-model release. Its purpose is to make the provenance of data, versions, assumptions, calculations, and engineering decisions reproducible and verifiable before a geological result is transferred to mine planning, mining, processing, or economic evaluation. The architecture must not, however, be interpreted as a mechanism for automatically removing defects in the empirical basis. The system intentionally does not ‘correct’ original sampling bias, orebody non-stationarity, sample-preparation errors, analytical uncertainty, laboratory-method bias, or the loss of spatial and temporal context unless new physical data enter the workflow. It can identify inconsistency, disclose uncertainty, block an unsupported release, require verification, or prioritise resampling. It may not substitute a statistical, generative, or agentic reconstruction for a physical observation. Accordingly, the practical effectiveness of the Geology Twin depends entirely on the quality of the initial empirical material and on the integrity of its evidence chain: sampling design and execution; sample representativity and preservation; chain of custody; preparation; analysis; certificates; standards, blanks, duplicates, and control assays; spatial referencing; density measurements; geological logging; and documented decisions to exclude or use data. A well-governed architecture does not convert a weak observation set into a reliable resource model; it makes the limits of that set visible and prevents their concealed propagation into downstream decisions. Quality control of the source material is established before system deployment, during the audit stage. The audit shall identify data owners and authoritative sources; test the completeness and traceability of the sampling and analytical chain; assess QA/QC, bias, repeats, gaps, domain and temporal non-stationarity; confirm the coordinate reference system, intervals, density data, compositing and exclusion rules; and establish the model’s permitted domain of use, uncertainty register, and stop conditions. Until critical non-conformities are closed, the architecture may be used only for observation, diagnosis, historical replay, and preparation of a programme for obtaining new data. It shall not be used to approve resource, production, or financial decisions. Mandatory implementation gate. The system does not correct original sampling bias, non-stationarity, sensor or survey bias, laboratory error, incomplete observation, unrecorded operational change, or any other defect embedded in the empirical basis. A technical review may identify, qualify, or reject a defect; it does not itself cure an empirical defect. Production or operational deployment, any controlled release, and any resource, planning, production, commercial, financial, safety, environmental, or regulatory decision are prohibited until an independent audit has successfully confirmed the quality of the empirical basis, QA/QC, provenance, stated limitations, and fitness of the evidence for the specific decision. Before that acceptance, permissible uses are limited to audit, observation, diagnosis, historical replay, shadow mode, and preparation of a programme to obtain new physical data. This limitation is a principle of engineering integrity, not a deficiency of Volume 1. The Geology Twin is intended to strengthen the professional judgement of the geologist and Competent Person, provide a transparent connection between data and conclusion, and direct field verification to locations where the value of new information is greatest. The release of a controlled geological model, a change in interpretation, a resource category, or a planning basis remains the responsibility of an authorised specialist acting under applicable reporting codes, corporate delegations of authority, and applicable law.