Three dimensional mineral prospectivity modeling by evidential belief functions, a case study from Kahang porphyry Cu deposit

Abstract

The present study has focused on the performance of the evidential belief functions (EBFs) model to predict the three dimensional (3D) mineralization of copper (Cu) at Kahang porphyry Cu deposit (KPCD) in the northeast of Isfahan, Iran. The 3D deposit modeling and reserve estimation are of the most important steps in mineral exploration program. Determining the boundaries of the deposit and estimating the quantity and quality of reserves are among the most essential issues in mining designing. Normally, a 3D model of the sought deposit is prepared based on the mineral grade parameter without considering other parameters and deposit information. In this regard, 3D integration models at the deposit scale facilitate the consideration of other parameters and available underground information. The results of these 3D integration models are mostly used to introduce mineral prospectivity areas. However, in this study, the integrated model was used as an auxiliary parameter in estimation for the geostatistical co-kriging method. Of note is that the accuracy of the 3D model of the deposit improved whereas the estimation error demonstrated a reduction. . The information of 22 boreholes drilled by the National Iranian Copper Industry Company was used in this research. To implement resource modeling, the number of blocks in the eastern, northern and depth directions were 30 × 28 × 145, respectively and the size of each block was designed equal to15 × 15 × 5 m. Auxiliary geospatial models were divided into three main categories of geological, geochemical, and geophysical data sets, and all 3D evidential models were obtained quantitatively. Subsequently, the Dempster-Shafer evidence theory was applied to construct a 3D prospectivity modeling of Kahang deposit. Using this knowledge-driven theory, four 3D models of belief, uncertainty, plausibility, and disbelief were produced, which have substantial impact on more accurate Cu resource modeling. After validation of the prospectivity models with the drilling results, the desired 3D auxiliary model had a correlation of 0.92 with the Cu grades. Thus, the Cu grade was estimated by a co-kriging method using a 3D EBFs model as an auxiliary variable, and output was compared with the one from an ordinary kriging (OK) method. The significance of this study lies in improvement of the performance of the Cu grade modeling through an auxiliary variable, where the root-mean-square error (RMSE) of the estimates reduces from 0.437 to 0.288 when implementing a co-kriging method rather than the ordinary kriging.