Answers are grounded in this paper and cite their sources.
Geophysical survey demonstrates effective targeting of porphyry copper deposits in BC, indicating machine learning can enhance mineral exploration.
The Huckleberry project and porphyry copper–molybdenum mine and adjoining Berg and Ootsa porphyry projects are located approximately 90 km southwest of Houston, British Columbia (Figure 1). Together they form a large and rich mineral endowment, with as many as six known porphyry deposits and numerous occurrences, making them ideally suited for study using airborne geophysical methods. In 2021, Geotech Ltd. carried out a helicopter-borne natural field electromagnetic (NFEM) and magnetic survey over the Berg–Ootsa and Huckleberry projects that consisted of a combined 5251 line-km of coverage. Analyses of the airborne geophysical responses and 3D inversion results show that the known porphyry deposits all coincide with well-defined subcircular or tabular resistivity lows and similar coincident magnetic high signatures but with varying degrees of size and amplitude. These geophysical signatures are consistent with those previously found in NFEM and other airborne electromagnetic surveys over other calc-alkaline type porphyry deposits in the Western Cordillera. Other similar geophysical signatures are observed, but efficient targeting of potential porphyries over the vast survey area proves difficult due to the large number and their variable nature. A semiautomated mineral targeting approach was implemented that uses a relatively objective, machine-learning-assisted method, which combines structural complexities, self-organizing map classifications, and supervised deep neural network (SDNN) targeting. Using the Huckleberry deposit as a training area, the SDNN targeting approach was extended to a larger area covering the Berg–Ootsa–Huckleberry projects and has identified most of the known porphyry deposits and prospects, as well as new areas for follow-up.
No takes yet. Share an insight, caveat, or question.
Legault et al. (2025) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: