Application of an ordinary kriging–artificial neural network for elemental distribution in Kahang porphyry deposit, Central Iran

• Published in: Arabian journal of geosciences Vol. 13; no. 15
• Main Authors: Yasrebi, Amir Bijan, Hezarkhani, Ardeshir, Afzal, Peyman, Karami, Reza, Tehrani, Mohammad Eskandarnejad, Borumandnia, Ali
• Format: Journal Article
• Language: English
• Published: Cham Springer International Publishing 01.08.2020; Springer Nature B.V
• Subjects: Artificial neural networks; Copper; Correlation coefficient; Correlation coefficients; Distribution; Earth and Environmental Science; Earth science; Earth Sciences; Eocene; Faults; Gold; Lithology; Mineral exploration; Molybdenum; Neural networks; Original Paper; Silver; Statistical methods; Volcanic rocks; Iran; Minimum error; Elemental distribution; Ordinary kriging–artificial neural network (OK-ANN); Kahang porphyry deposit
• ISSN: 1866-7511; 1866-7538
• DOI: 10.1007/s12517-020-05607-0

Estimation of elemental distribution with minimum estimation errors is crucial in mineral exploration. This study aims to estimate Cu, Mo, Ag, and Au distribution in a local scale. To do this, the application of an ordinary kriging–artificial neural network (OK-ANN) model is proposed for elemental distribution based on lithogeochemical data in Kahang porphyry deposit, Central Iran. In addition, ordinary kriging (OK) method is applied and the result will be compared with the suggested method. Following this, the proposed estimation model has less error in comparison with the OK. To ensure the validity of proposed OK-ANN, the results were correlated with lithological, faults, and alteration maps in the studied region using a logratio matrix. As a result, main parts of Ag and Au enrichment are associated with faults and argillic alteration with high values of overall accuracies (OAs) in the western part of the deposit. Furthermore, Cu values estimated by the OK-ANN are connected with potassic alteration zone in the central part of deposit. In addition, the Cu-, Mo-, Au-, and Ag-enriched anomalous parts are associated with the existing faults, major alterations, and Eocene sub-volcanic rocks as well as porphyritic quartz diorites, monzodiorite–monzogranite, and dacitic units. Correlation coefficient between raw data and the results obtained by the proposed estimator is more than 84%, meaning that the proposed technique can be utilized for estimation of geochemical data in porphyry deposits.