Three-dimensional weights of evidence modeling of a deep-seated porphyry Cu deposit

TL;DR

This paper introduces a 3D weights of evidence modeling framework for deep-seated porphyry copper deposits, integrating geological and geochemical data to improve prospectivity predictions.

Contribution

It presents a novel 3D WofE-based approach combining qualitative and quantitative data, validated with the Nochoun deposit and provided as open-source software.

Findings

Studentized posterior probability model outperforms in identifying copper mineralized zones.

Fuzzy weights effectively incorporate continuous geochemical data.

The framework enhances deep-seated ore body exploration.

Abstract

Given the challenges in data acquisition and modeling at the stage of detailed exploration, it is difficult to develop a prospectivity model, particularly for disseminated ore deposits. Recently, the weights of evidence (WofE) method has demonstrated a high efficiency for modeling such deposits. In this study, we propose a framework for creating a three-dimensional (3D) weights of evidence-based prospectivity model of the Nochoun porphyry Cu deposit in the Urmia-Dokhtar magmatic arc of Iran. The input data include qualitative geological and quantitative geochemical information obtained from boreholes and field observations. We combine ordinary and fuzzy weights of evidence for integrating qualitative and quantitative exploration criteria in a 3D space constrained by a metallogenic model of the study area for identifying a deep-seated ore body. Ordinary weights of evidence are determined for geological data, including lithology, alteration, rock type, and structure. Moreover, we determine the fuzzy weight of evidence for each class of continuous geochemical models created based on Fe, Mo, and Zn concentration values derived from boreholes. We integrate the input evidential models using WofE and create two prospectivity models (i.e., posterior and studentized posterior probability). We also determine anomalous voxels in the probability models using concentration-volume fractal models and validate them using prediction-volume plots. The modeling results indicate that the studentized posterior probability model is more efficient in identifying voxels representing copper mineralized rock volumes. We provide open source software for the proposed framework which can be used for exploring deep-seated ore bodies in other regions.