Machine learning of pyrite geochemistry reconstructs the multi-stage history of mineral deposits

Abstract

The application of machine learning for pyrite discrimination establishes a robust foundation for constructing the ore-forming history of multi-stage deposits; however, published models face challenges related to limited, imbalanced datasets and oversampling. In this study, the dataset was expanded to approximately 500 samples for each type, including 508 sedimentary, 573 orogenic gold, 548 sedimentary exhalative (SEDEX) deposits, and 364 volcanogenic massive sulfides (VMS) pyrites, utilizing random forest (RF) and support vector machine (SVM) methodologies to enhance the reliability of the classifier models. The RF classifier achieved an overall accuracy of 99.8%, and the SVM classifier attained an overall accuracy of 100%. The model was evaluated by a five-fold cross-validation approach with 93.8% accuracy for the RF and 94.9% for the SVM classifier. These results demonstrate the strong feasibility of pyrite classification, supported by a relatively large, balanced dataset and high accuracy rates. The classifier was employed to reveal the genesis of the controversial Keketale Pb-Zn deposit in NW China, which has been inconclusive among SEDEX, VMS, or a SEDEX-VMS transition. Petrographic investigations indicated that the deposit comprises early fine-grained layered pyrite (Py1) and late recrystallized pyrite (Py2). The majority voting classified Py1 as the VMS type, with an accuracy of RF and SVM being 72.2% and 75%, respectively, and confirmed Py2 as an orogenic type with 74.3% and 77.1% accuracy, respectively. The new findings indicated that the Keketale deposit originated from a submarine VMS mineralization system, followed by late orogenic-type overprinting of metamorphism and deformation, which is consistent with the geological and geochemical observations. This study further emphasizes the advantages of Machine learning (ML) methods in accurately and directly discriminating the deposit types and reconstructing the formation history of multi-stage deposits.

Keywords

Machine learning / Random forest / Support vector machine / Pyrite / Multi-stage genesis / Keketale deposit

CRediT authorship contribution statement

Pengpeng Yu: Writing – review & editing, Writing – original draft, Supervision, Methodology, Conceptualization. Yuan Liu: Writing – original draft, Investigation, Formal analysis, Data curation. Hanyu Wang: Visualization, Validation, Investigation. Xi Chen: Writing – review & editing, Supervision, Methodology, Conceptualization. Yi Zheng: Supervision, Project administration, Funding acquisition. Wei Cao: Investigation. Yiqu Xiong: Writing – review & editing, Supervision. Hongxiang Shan: Investigation.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgments

This research is jointly funded by the National Key Research and Development Program of China (2021YFC2900300), the Natural Science Foundation of Guangdong Province (2024A1515030216), MOST Special Fund from State Key Laboratory of Geological Processes and Mineral Resources, China University of Geosciences (GPMR202437), the Guangdong Province Introduced of Innovative R&D Team (2021ZT09H399) and the Third Xinjiang Scientific Expedition Program (2022xjkk1301). We appreciate the editor-in-chief and anonymous reviewers for their constructive feedback and comments.

References

[1] Z.U. Bajwah, P.K. Seccombe, R. Offler. Trace element distribution Co: Ni ratios and genesis of the big Cadia iron-copper deposit, New South Wales, Australia. Mineral. Deposita, 22 (1987), pp. 292-300,

[2] I. Belousov, R.R. Large, S. Meffre, L.V. Danyushevsky, J. Steadman, T. Beardsmore. Pyrite compositions from VHMS and orogenic Au deposits in the Yilgarn Craton, Western Australia: Implications for gold and copper exploration. Ore Geol. Rev., 79 (2016), pp. 474-499,

[3] Breiman, L., 2001. Random forests. Mach. Learn. 45, 5–32. https://doi.org/10.1023/A:1010933404324.

[4] L. Breiman. Bagging predictors. Mach. Learn., 24 (1996), pp. 123-140,

[5] B.A. Brill. Trace-element contents and partitioning of elements in ore minerals from the CSA Cu-Pb-Zn deposit, Australia, and implications for ore genesis. Can. Mineral., 27 (7) (1989), pp. 263-274

[6] A. Broussolle, M. Sun, K. Schulmann, A. Guy, C. Aguilar, P. Štípská, Y. Jiang, Y. Yu, W. Xiao. Are the Chinese Altai “terranes” the result of the juxtaposition of different crustal levels during Late Devonian and Permian orogenesis?. Gondwana Res., 66 (2019), pp. 183-206,

[7] F. Chai. Geochronology and Genesis of Meta-felsic Volcanic Rocks from the Kangbutiebao Formation in Chonghuer Basin on Southern Margin of Altay, Xinjiang. Geol. Rev., 58 (6) (2012), pp. 1023-1037

[8] F. Chai, J. Mao, L. Dong, F. Yang, F. Liu, X. Geng, Z. Zhang. Geochronology of metarhyolites from the Kangbutiebao Formation in the Kelang basin, Altay Mountains, Xinjiang: Implications for the tectonic evolution and metallogeny. Gondwana Res., 16 (2009), pp. 189-200,

[9] Chen, Y.J., Pirajno, F., Wu, G., Qi, J.P., Xiong, X.L., 2012. Epithermal deposits in North Xinjiang, NW China. Int. J. Earth Sci. 101, 889–917. https://doi.org/

[10] N.J. Cook. Mineralogy of the sulphide deposits at Sulitjelma, northern Norway. Ore Geol. Rev., 11 (1996), pp. 303-338,

[11] C. Cortes, V. Vapnik. Support-vector networks. Mach. Learn., 20 (1995), pp. 273-297,

[12] H. Cui, Y. Deng, R. Zhong, W. Li, C. Yu, L.V. Danyushevsky, I. Belousov, Z. Li, H. Wang. Determining the ore-forming processes of Dongshengmiao Zn-Pb-Cu deposit: Evidence from the linear discriminant analysis of pyrite geochemistry. Ore Geol. Rev., 163 (2023), Article 105782,

[13] Danyushevsky, L., Robinson, P., Gilbert, S., Norman, M., Large, R., McGoldrick, P., Shelley, M., 2011. Routine quantitative multi-element analysis of sulphide minerals by laser ablation ICP-MS: Standard development and consideration of matrix effects. Geochem. Explor. Environ. Anal. 11, 51–60. https://doi.org/

[14] Gadd, M.G., Layton-Matthews, D., Peter, J.M., Paradis, S.J., 2016. The world-class Howard’s Pass SEDEX Zn-Pb district, Selwyn Basin, Yukon. Part I: trace element compositions of pyrite record input of hydrothermal, diagenetic, and metamorphic fluids to mineralization. Mineral. Deposita 51, 319–342. https://doi.org/

[15] R. Goldfarb, J. Mao, C. Hart, D. Wang, E. Anderson, Z. Wang. Tectonic and metallogenic evolution of the Altay Shan, Northern Xinjiang Uygur Autonomous region, Northwestern China. Tectonic Evolution and Metallogeny of the Chinese Altay and Tianshan: London, IAGOD Guidebook Ser., 10 (2003), pp. 17-30,

[16] R.J. Goldfarb, R.D. Taylor, G.S. Collins, N.A. Goryachev, O.F. Orlandini. Phanerozoic continental growth and gold metallogeny of Asia. Gondwana Res., 25 (2014), pp. 48-102,

[17] A.D. Gordon, L. Breiman, J.H. Friedman, R.A. Olshen, C.J. Stone. Classification and regression trees. Biometrics, 40 (1984), p. 874,

[18] D.D. Gregory, M.J. Cracknell, R.R. Large, P. McGoldrick, S. Kuhn, V.V. Maslennikov, M.J. Baker, N. Fox, I. Belousov, M.C. Figueroa, J.A. Steadman, A.J. Fabris, T.W. Lyons. Distinguishing ore deposit type and barren sedimentary pyrite using laser ablation-inductively coupled plasma-mass spectrometry trace element data and statistical analysis of large data sets. Econ. Geol., 114 (2019), pp. 771-786,

[19] D.D. Gregory, R.R. Large, J.A. Halpin, E.L. Baturina, T.W. Lyons, S. Wu, L. Danyushevsky, P.J. Sack, A. Chappaz, V.V. Maslennikov, S.W. Bull. Trace element content of sedimentary pyrite in black shales. Econ. Geol., 110 (2015), pp. 1389-1410,

[20] D.I. Groves, R.J. Goldfarb, M. Santosh. The conjunction of factors that lead to formation of giant gold provinces and deposits in non-arc settings. Geosci. Front., 7 (2016), pp. 303-314,

[21] D.I. Groves, M. Santosh, R.J. Goldfarb, L. Zhang. Structural geometry of orogenic gold deposits: Implications for exploration of world-class and giant deposits. Geosci. Front., 9 (2018), pp. 1163-1177,

[22] Hannington, M.D., 2014. Volcanogenic Massive Sulfide Deposits. In: Holland, H.D., Turekian, K.K. (Eds.), Treatise on Geochemistry (Second Edition). Elsevier, p. 463–488. https://doi.org/10.1016/B978-0-08-095975-7.01120-7.

[23] J. Jiang. Genesis and wallrock alteration of the Keketale Pb-Zn deposit in Aertai, Xinjiang. Mineral Resour. Geol., 17 (6) (2003), pp. 679-682

[24] K. Jochum, U. Weis, B. Stoll, D. Kuzmin, D. Jacob, A. Stracke, K. Birbaum, D. Frick, D. Günther, J. Enzweiler. Determination of reference values for NIST SRM 610–617 glasses following ISO Guidelines. Geostand. Geoanal. Res., 35 (2011), pp. 397-429,

[25] J.D. Kelleher, N.B. Mac, A. D’arcy. Fundamentals of Machine Learning for Predictive Data Analytics: Algorithms, Worked Examples, and Case Studies. MIT Press, Cambridge, MA (2015)

[26] R.R. Large, S.W. Bull, V.V. Maslennikov. A carbonaceous sedimentary source-rock model for carlin-type and orogenic gold deposits. Econ. Geol., 106 (2011), pp. 331-358,

[27] D.L. Leach, D.C. Bradley, D. Huston, S.A. Pisarevsky, R.D. Taylor, S.J. Gardoll. Sediment-hosted lead-zinc deposits in Earth history. Econ. Geol., 105 (3) (2010), pp. 593-625,

[28] H.X. Li, Z.C. Zhang, R.X. Zhang, Q.H. Xie, L. Zhang, M. Santosh. Geochemical discrimination of pyrite in diverse ore deposit types through statistical analysis and machine learning techniques. Am. Mineral., 109 (5) (2024), pp. 846-857,

[29] P. Li, C. Yuan, M. Sun, X. Long, K. Cai. Thermochronological constraints on the late Paleozoic tectonic evolution of the southern Chinese Altai. J. Asian Earth Sci., 113 (2015), pp. 51-60,

[30] R. Li, H. Chen, R.R. Large, L. Zhao, Y. Liu, J. Jiao, X.-P. Xia, Q. Yang. Ore-forming fluid source of the orogenic gold deposit: Implications from a combined pyrite texture and geochemistry study. Chem. Geol., 552 (2020), Article 119781,

[31] X.M. Li, Y.X. Zhang, Z.K. Li, X.F. Zhao, R.G. Zuo, F. Xiao, Y. Zheng. Discrimination of Pb-Zn deposit types using sphalerite geochemistry: New insights from machine learning algorithm. Geosci. Front., 14 (2023), Article 101580,

[32] J.J. Lindsay, H.S.R. Hughes, C.M. Yeomans, J.C.Ø. Andersen, I. McDonald. A machine learning approach for regional geochemical data: Platinum-group element geochemistry vs geodynamic settings of the North Atlantic Igneous Province. Geosci. Front., 12 (2021), Article 101098,

[33] H.M. Liu, G. Beaudoin, S. Makvandi, S.E. Jackson, X.W. Huang. Multivariate statistical analysis of trace element compositions of native gold from orogenic gold deposits: Implication for mineral exploration. Ore Geol. Rev., 131 (2021), Article 104061,

[34] Y. Liu, Z. Hu, S. Gao, D. Günther, J. Xu, C. Gao, H. Chen. In situ analysis of major and trace elements of anhydrous minerals by LA-ICP-MS without applying an internal standard. Chem. Geol., 257 (2008), pp. 34-43,

[35] Y. Liu, P.P. Yu, Y. Zheng, H.X. Shan, X. Chen. Texture and composition evolution of sphalerite in metamorphosed deposits: An example from the Keketale Pb-Zn(-Ag) deposit, NW China. Ore Geol. Rev., 165 (2024), Article 105924,

[36] X. Long, M. Sun, C. Yuan, W. Xiao, K. Cai. Early Paleozoic sedimentary record of the Chinese Altai: Implications for its tectonic evolution. Sediment Geol., 208 (2008), pp. 88-100,

[37] J.W. Mao, F. Pirajno, Z.H. Zhang, F.M. Chai, H. Wu, S.P. Chen, L.S. Cheng, J.M. Yang, C.Q. Zhang. A review of the Cu–Ni sulphide deposits in the Chinese Tianshan and Altay orogens (Xinjiang Autonomous Region, NW China): Principal characteristics and ore-forming processes. J. Asian Earth Sci., 32 (2008), pp. 184-203,

[38] A. Nadi, H. Moradi. Increasing the views and reducing the depth in random forest. Exp. Syst. Appl., 138 (2019), Article 112801.,

[39] C.L. Nathwani, J.J. Wilkinson, G. Fry, R.N. Armstrong, D.J. Smith, C. Ihlenfeld. Machine learning for geochemical exploration: classifying metallogenic fertility in arc magmas and insights into porphyry copper deposit formation. Mineral. Deposita, 57 (2022), pp. 1143-1166,

[40] G. O’Sullivan, D. Chew, G. Kenny, I. Henrichs, D. Mulligan. The trace element composition of apatite and its application to detrital provenance studies. Earth Sci. Rev., 201 (2020), Article 103044,

[41] C. Paton, J. Hellstrom, B. Paul, J. Woodhead, J. Hergt. Iolite: Freeware for the visualisation and processing of mass spectrometric data. J. Anal. at. Spectrom., 26 (12) (2011), pp. 2508-2518,

[42] M. Petrelli. Machine learning in petrology: State-of-the-art and future perspectives. J. Petrol., 65 (5) (2024), Article egae036,

[43] F. Pirajno, M. Santosh. Rifting, intraplate magmatism, mineral systems and mantle dynamics in central-east Eurasia: An overview. Ore Geol. Rev., 63 (2014), pp. 265-295,

[44] K.Z. Qin, J.B. Wang, J.H. Zhang, J.N. Deng. Metallogenic conditions for the Keketale-style large-scale Pb-Zn deposit on the southern margin of Altay, Xinjiang. Geol. Explor. Non-Ferrous Metals., 7 (2) (1998), pp. 65-74

[45] Qu, H.Y., Mao, J.W., Zhou, S.M., 2020. Metallogenesis of stratiform Cu mineralization in the Dabaoshan polymetallic deposit, Northern Guangdong Province, South China. J. Geochem. Explor. 210, 106448. https://doi.org/

[46] M. Reich, S.E. Kesler, S. Utsunomiya, C.S. Palenik, S.L. Chryssoulis, R.C. Ewing. Solubility of Au in arsenian pyrite. Geochim. Cosmochim. Acta, 69 (2005), pp. 2781-2796,

[47] I.H. Sarker. Machine learning: Algorithms, real-world applications and research directions. SN Comput. Sci., 2 (2021), p. 160,

[48] C. Sun, X. Yang, H. Zhang, W. Ji, B. Chen, Z. Dong, M. Faisal, D. Xi. Tracing the formation and modification of the Keketale VMS-type Pb-Zn deposit, Altai Mountains: Insights from ore deposit geology, geochronology, and magnetite geochemistry. Ore Geol. Rev., 144 (2022), Article 104852,

[49] R.C. Tan, Y.J. Shao, M.J. Brzozowski, Y. Zheng, Y.Q. Xiong. Development of a machine learning model to classify mineral deposits using sphalerite chemistry and mineral assemblages. Ore Geol. Rev., 169 (2024), Article 106076,

[50] Thomas, C.W., Aitchison, J., 2006. Log-ratios and geochemical discrimination of Scottish Dalradian limestones: a case study. J. Geol. Soc. London 264, 25–41. https://doi.org/

[51] N. Tribovillard, T.J. Algeo, T. Lyons, A. Riboulleau. Trace metals as paleoredox and paleoproductivity proxies: an update. Chem. Geol., 232 (2006), pp. 12-32,

[52] K. Ueki, H. Hino, T. Kuwatani. Geochemical discrimination and characteristics of magmatic tectonic settings: A machine-learning-based approach. Geochem. Geophys. Geosyst., 19 (2018), pp. 1327-1347,

[53] J.B. Wang, K.Z. Qin, Z.L. Wu, J.H. Hu, J.N. Deng. Volcanic-exhalative-sedimentary Lead Zinc Deposit in the Southern Margin of the Altai, Xinjiang. Geology Publishing House, Beijing (1998)

[54] J.B. Wang, J.N. Deng, J.H. Zhang, K.Z. Qin. Massive sulphide deposits related to the volcano-passive continental margin in the Altay Region. Acta Geol. Sin., 73 (2) (1999), pp. 253-263,

[55] S.L. Wang, Z.L. Guo, Y.W. Wang, Z.L. Mao. Geological characteristics of Pb-Zn deposits in Devonian volcanic-sedimentary basins in the south margin of Altay Mountain: a case study of Keketale Pb-Zn deposit, Xinjiang, China. Geol. Explor., 6 (2005), pp. 27-33

[56] S.L. Wang, K.Q. Chen, K.C. Kang, Q. Guo. Stable isotope of Pb-Zn deposits occurred in the Maizi Devonian volcanic sedimentary basin in the south margin of Altay mountain, Xinjiang. Geol. Explor., 6 (2007), pp. 25-31

[57] Y. Wang, K.F. Qiu, A. Müller, Z.L. Hou, Z.H. Zhu, H.C. Yu. Machine learning prediction of Quartz forming-environments. J. Geophys. Res. Solid Earth, 126 (2021), Article e2021JB021925,

[58] R. Wei, J. Wang, M. Su, E. Jia, S. Chen, T. Chen, Y. Ni. Missing value imputation approach for mass spectrometry-based metabolomics data. Sci. Rep., 8 (2018), p. 663,

[59] C. Wilhem, B. Windley, G. Stampfli. The Altaids of Central Asia: A tectonic and evolutionary innovative review. Earth Sci. Rev., 113 (2012), pp. 303-341,

[60] Wilson, S., Ridley, I., Koenig, A., 2002. Development of sulfide calibration standards for the laser ablation inductively-coupled plasma mass spectrometry technique US Government. J. Anal. Atom Spectrom. 17, 406–409. https://doi.org/

[61] B. Windley, D. Alexeiev, W. Xiao, A. Kröner, G. Badarch. Tectonic models for accretion of the Central Asian Orogenic Belt. J. Geol. Soc., 164 (2007), pp. 31-47,

[62] Z. Wu, Z. Wang, J.J. Chen, H.M. You, M. Yan, L.J. Wang. Stratified random sampling for neural network test input selection. Inform. Software Tech., 165 (2024), Article 107331,

[63] W. Xiao, M. Santosh. The western Central Asian Orogenic Belt: A window to accretionary orogenesis and continental growth. Gondwana Res., 25 (2014), pp. 1429-1444,

[64] C. Yang, F. Yang, F. Chai, Y. Wu. Timing of formation of the Keketale Pb–Zn deposit, Xinjiang, Northwest China, Central Asian Orogenic Belt: Implications for the metallogeny of the South Altay Orogenic Belt. Geol. J., 53 (2018), pp. 899-913,

[65] K. Yao, Y. Zheng. Nanophotonics and Machine Learning: Concepts, Fundamentals, and Applications. Springer International Publishing, Cham (2023), pp. 77-112

[66] L. Ye, N.J. Cook, C.L. Ciobanu, Y.P.L. Liu, Q. Zhang, T.G. Liu, W. Guo, Y.L. Yang, L. Danyushevskiy. Trace and minor elements in sphalerite from base metal deposits in South China: A LA-ICPMS study. Ore Geol Rev., 39 (2011), pp. 188-217,

[67] P. Yu, Y. Zheng. Pb-Zn-Cu accumulation from seafloor sedimentation to metamorphism: Constraints from ore textures coupled with elemental and isotopic geochemistry of the Tiemurt in Chinese Altay Orogen, NW China. Gondwana Res., 72 (2019), pp. 65-82,

[68] P. Yu, Y. Zheng, C.M. Wang. Trace elemental and sulfur-lead isotopic variations in metamorphosed volcanogenic massive sulfide (VMS) mineralization systems: An example from the Keketale Pb-Zn(-Ag) deposit, NW China. Ore Geol Rev., 125 (2020), Article 103685,

[69] J. Zhang, M. Sun, K. Schulmann, G. Zhao, Q. Wu, Y. Jiang, A. Guy, Y. Wang. Distinct deformational history of two contrasting tectonic domains in the Chinese Altai: Their significance in understanding accretionary orogenic process. J. Struct. Geol., 73 (2015), pp. 64-82,

[70] P. Zhang, Z. Zhang, J. Yang, Q. Cheng. Machine learning prediction of ore deposit genetic type using magnetite geochemistry. Nat. Resour. Res., 32 (2023), pp. 99-116,

[71] Zheng, Y., Zhang, L., Chen, H., Li, D., Wang, C., Fang, J., 2014. CO2-rich fluid from metamorphic devolatilization of the Triassic Orogeny: an example from the Qiaxia copper deposit in Altay, NW China. Geol. J. 49, 617–634. https://doi.org/

[72] Y. Zheng, L. Zhang, P. Hollings, H. Chen. Metamorphosed Pb–Zn–(Ag) ores of the Keketale VMS deposit, NW China: Evidence from ore textures, fluid inclusions, geochronology and pyrite compositions. Ore Geol. Rev., 54 (2013), pp. 167-180,

[73] Y. Zheng, L. Zhang, D.F. Li, A. Kapsiotis, Y.J. Chen. Genesis of the Dadonggou Pb–Zn deposit in Kelan basin, Altay, NW China: Constraints from zircon U–Pb and biotite 40Ar/39Ar geochronological data. Ore Geol. Rev., 64 (2015), pp. 128-139,

[74] R. Zhong, Y. Deng, W. Li, L.V. Danyushevsky, M.J. Cracknell, I. Belousov, Y. Chen, L. Li. Revealing the multi-stage ore-forming history of a mineral deposit using pyrite geochemistry and machine learning-based data interpretation. Ore Geol. Rev., 133 (2021), Article 104079,

[75] Zou, S., Chen, X., Brzozowski, M.J., Leng, C.B., Xu, D., 2022. Application of machine learning to characterizing magma fertility in porphyry Cu deposits. J. Geophys. Res. Solid Earth 127, e2022JB024584. https://doi.org/