125 We intend to show a method that introduces a quantitative metric of spatial similarity based on spatial principal component analysis (sPCA). By decomposing into spatial eigenfrequencies, sPCA captures the specific geophysical signature of the reference deposit and quantifies how closely other regions reproduce that structure. The aim is to enable the detection of deposit-specific spatial patterns without labeled data or filters. Beyond being fully data-driven, the approach is designed to complement and enhance expert-driven exploration workflows. It seeks to enable quantitative pattern recognition that can sit atop existing geological knowledge, allowing experts to iteratively refine or validate hypotheses through an interpretable, datainformed similarity metric. We intend to show this approach in a brownfield case study in the Carajás Mineral Province, Brazil, applying it to magnetic and radiometric datasets to identify IOCG-related anomalies consistent with a known copper deposit. 2. STUDY AREA AND DATA The study area lies in the Carajás Mineral Province of Pará, northern Brazil, an Archean granite– greenstone and volcano-sedimentary terrane of the Itacaiúnas Supergroup that hosts numerous Fe-oxide and IOCG systems (Error! Reference source not found.). The aim is to detect locations within this region whose internal spatial patterns resemble those expressed by the geophysical signature of a known IOCG deposit and to rank them by similarity. Case 1 uses the Paulo Afonso deposit (Dep. 6;Figure) as the univariate reference. Paulo Afonso is an iron-oxide copper–gold (IOCG) body characterized by magnetite–hematite–rich alteration and structurally focused chalcopyrite Au mineralization that produces a strong dipole anomaly in total magnetic intensity (TMI). This TMI dipole response is a diagnostic signature that IOCG experts already use as an exploration criterion in this area, making it a geologically meaningful template for identifying district-scale analogs and assessing regional prospectivity. To enable a similar-to-similar comparison between the deposit and the surrounding geographic area, the TMI grid is analyzed using a sliding-window approach. Every window has the same dimensions as the deposit footprint. The resulting analysis window measures 46 × 28 pixels, corresponding to a physical area of 9.20 × 5.60 km at the 200-m grid resolution, and is moved across the full 259 × 138 km demonstration area. Within this corridor, several other economic IOCG and iron-ore deposits exhibit comparable magnetic dipoles, providing a natural success metric for validating the predictive method and ranking targets.
RkJQdWJsaXNoZXIy MTM0Mzk2