Fitzroy Minerals Inc. has completed an integrated seismic and machine-learning targeting study with Australian technology firm Fleet Space Technologies at its Buen Retiro Copper Project near Copiapo, Chile, generating a ranked portfolio of nine exploration targets that have never been drilled. The announcement, reported via NEWMEDIAWIRE, marks an effort to apply advanced subsurface imaging and data science to a large iron oxide copper-gold system where extensive gravel cover has historically hampered conventional target generation.
The machine-learning model combined Fleet’s 3D Ambient Noise Tomography data with Fitzroy’s own geophysical surveys and approximately 24,400 copper assays from 140 drill holes. It identified nine areas where subsurface characteristics closely resemble the signature of rocks hosting known copper mineralization at Buen Retiro. All targets lie within 400 metres of surface and exclude previously mined ground or areas within 200 metres of an existing hole. A summary of the four datasets and assay results used is available in Table 1.
The significance lies in what the targets represent: blind exploration opportunities beneath alluvial gravel that has made surface mapping difficult. The ANT surveys mapped the project’s deeper geology well beyond the depths of current drilling, giving Fitzroy a 3D framework to trace known mineralized trends and identify new opportunities, including toward the 100%-owned Sierra Fritis properties to the east. Fitzroy has started a 5,000-metre drilling program using diamond and reverse circulation rigs.
Merlin Marr-Johnson, President and CEO of Fitzroy, said the company believes Buen Retiro shares many geological characteristics of major IOCG deposits at Candelaria and Mantoverde, yet the sulphide potential of the wider property has barely been tested. “Now, however, we have a tool with which to test the exploration optionality of this large IOCG system in one of Chile’s most prolific copper-producing regions,” he said. “If the ML targeting is successful it will be a cost-effective discovery tool, creating value for shareholders.”
The targeting method followed four steps: prediction, in which the model learned relationships between geophysical data and copper assays to estimate zones where grades above a 0.5% copper threshold might occur; refinement, removing blocks within the mined-out pit, mapped gravel cover, or within 200 metres of existing holes; grouping remaining blocks into nine geographic domains; and drill planning. As drill logs and assay results arrive, Fitzroy and Fleet will update the model, so each new hole tests a target, validates the model, and informs the next drilling decision.
Fitzroy cautions that the domains are conceptual exploration targets not yet tested by drilling and do not confirm the presence, grade, size or continuity of mineralization. The reliability of the model depends on data quality and coverage and tends to decrease with distance from existing drill holes and with depth. The targets are intended to complement, not replace, conventional geological interpretation. Dr. Scott Jobin-Bevans, a Qualified Person under NI 43-101, has reviewed and approved the technical information.
The initiative matters because it tests whether machine learning can convert a mature drilling database and seismic imaging into cost-effective discovery opportunities in a prolific copper region, while the simultaneous 5,000-metre program will provide the ground truth needed to judge the approach.
