Prediction of copper content in waste dump of Sarcheshmeh copper mine using visible and near‑infrared reflectance spectroscopy

Abstract

In Sarcheshmeh, the amount of waste rock resulting from extraction and processing is very high due to the low grade of copper in host rocks. Weathering and oxidation of sulfide minerals increase the mobility and transfer rate of toxic elements in these environments. Classic sampling and laboratory analyses method should be performed to evaluate the source, type, and extent of the contamination. Given the disadvantages of these classical measurement methods, a quick and inexpensive way is required to monitor the amount of elements. Visible, near, and shortwave infrared reflectance (VNIR/SWIR) spectroscopy was used in this study, to predict the copper level in one of the waste dumps of Sarcheshmeh copper mine. 120 waste samples were gathered, and various experiments performed to determine the mineralogy and content of toxic elements in them. The reflectance spectra of the samples were also measured in the laboratory by means of a portable spectrometer. After preprocessing of the raw spectra, prediction mechanism was studied and prediction models were developed. Comparing to the iron content of samples, clay minerals had more significant impact on prediction of the copper. The best and worst prediction performances were obtained by stepwise multiple linear regression (SMLR) ( R2 p = 0.89) and principal component regression (PCR) ( R2 p = 0.37) methods, respectively. In addition, an SMLR predicted copper map, had relatively acceptable spatial similarity with the map constructed using the measured copper values. Results of this study showed that VNIR/SWIR spectroscopy can be considered as a novel method for predicting the copper content in highly heterogeneous environments.