Leucogranite mapping via convolutional recurrent neural networks and geochemical survey data in the Himalayan orogen

Abstract

Geochemical survey data analysis is recognized as an implemented and feasible way for lithological mapping to assist mineral exploration. With respect to available approaches, recent methodological advances have focused on deep learning algorithms which provide access to learn and extract information directly from geochemical survey data through multi-level networks and outputting end-to-end classification. Accordingly, this study developed a lithological mapping framework with the joint application of a convolutional neural network (CNN) and a long short-term memory (LSTM). The CNN-LSTM model is dominant in correlation extraction from CNN layers and coupling interaction learning from LSTM layers. This hybrid approach was demonstrated by mapping leucogranites in the Himalayan orogen based on stream sediment geochemical survey data, where the targeted leucogranite was expected to be potential resources of rare metals such as Li, Be, and W mineralization. Three comparative case studies were carried out from both visual and quantitative perspectives to illustrate the superiority of the proposed model. A guided spatial distribution map of leucogranites in the Himalayan orogen, divided into high-, moderate-, and low-potential areas, was delineated by the success rate curve, which further improves the efficiency for identifying unmapped leucogranites through geological mapping. In light of these results, this study provides an alternative solution for lithologic mapping using geochemical survey data at a regional scale and reduces the risk for decision making associated with mineral exploration.

Keywords

Lithological mapping / Deep learning / Convolutional neural network / Long short-term memory / Leucogranites

References

[1] Agterberg F.P., Bonham-Carter G.F., 2005. Measuring the performance of mineral-potential maps. Nat. Resour. Res. 14, 1-17.

[2] Altman N., Krzywinski M., 2017. Ensemble methods: bagging and random forests. Nat. Methods 14, 933-935.

[3] Ardabili S., Mosavi A., Várkonyi-Kóczy A.R., 2019. Advances in machine learning modeling reviewing hybrid and ensemble methods. International Conference on Global Research and Education 215-227.

[4] Breiman L., 2001. Random forests. Machine Learning 45, 5-32.

[5] Cai H., Chen T., 2020. Multi-Dimension CNN for Hyperspectral Image Classification. IEEE International Geoscience and Remote Sensing Symposium 1275-1278.

[6] Cao H., Li G., Zhang Z., Zhang L., Dong S., Xia X., Liang W., Fu J., Huang Y., Xiang A., 2020. Miocene Sn polymetallic mineralization in the Tethyan Himalaya, southeastern Tibet: A case study of the Cuonadong deposit. Ore Geo. Rev. 119, 103403.

[7] Cao H., Pei Q., Santosh M., Li G., Zhang L., Zhang X., Zhang Y., Zou H., Dai Z., Lin B. Tang L., 2022. Himalayan leucogranites: A review of geochemical and isotopic characteristics, timing of formation, genesis, and rare metal mineralization. Earth-Sci. Rev. 234, 104229.

[8] Carranza E.J.M., Hale M., Faassen C., 2008. Selection of coherent deposit-type locations and their application in data-driven mineral prospectivity mapping. Ore Geol. Rev. 33, 536-558.

[9] Cheng Q., Bonham-Carter G., Wang W., Zhang S., Li W., Xia Q., 2011. A spatially weighted principal component analysis for multi-element geochemical data for mapping locations of felsic intrusions in the Gejiu mineral district of Yunnan, China. Comput. Geosci. 37, 662-669.

[10] Cheng L., Zhang C., Liu X., Yang X., Zhou Y., Horn I., Weyer S., Holtz F., 2021. Significant boron isotopic fractionation in the magmatic evolution of Himalayan leucogranite recorded in multiple generations of tourmaline. Chem. Geol. 571, 120194.

[11] Cracknell M.J., Reading A.M., McNeill A.W., 2014. Mapping geology and volcanic-hosted massive sulfide alteration in the Hellyer-Mt Charter region, Tasmania, using Random Forests^TM and Self-Organising Maps. Aust. j.Earth Sci. 61, 287-304.

[12] Daneshfar B., Cameron E., 1998. Leveling geochemical data between map sheets. j.Geochem. Explor. 63, 189-201.

[13] Fawcett T., 2006. An introduction to ROC analysis. Pattern Recogn. Lett. 27, 861-874.

[14] Feurer M., Hutter F., 2019. Hyperparameter optimization. Automated Machine Learning: Methods, Systems, Challenges, 3-33.

[15] Gao L., Zeng L., Asimow P.D., 2017. Contrasting geochemical signatures of fluid-absent versus fluid-fluxed melting of muscovite in metasedimentary sources: the Himalayan leucogranites. Geology 45, 39-42.

[16] Gorgolis N., Hatzilygeroudis I., Istenes Z., Gyenne L.G., 2019. Hyperparameter optimization of LSTM network models through genetic algorithm. IEEE 10th International Conference on Information, Intelligence, Systems and Applications (IISA) 1-4.

[17] Graves A., 2012. Long short-term memory. Supervised sequence labelling with recurrent neural networks37-45.

[18] Greff K., Srivastava R.K., Koutník J., Steunebrink B.R., Schmidhuber J., 2016. LSTM: A search space odyssey. IEEE T. Neur. Net. Lear. Sys. 28, 2222-2232.

[19] Grunsky E.C., Mueller U.A., Corrigan D., 2014. A study of the lake sediment geochemistry of the Melville Peninsula using multivariate methods: applications for predictive geological mapping. j.Geochem. Explor. 141, 15-41.

[20] Guartán J.A., Emery X., 2021. Regionalized classification of geochemical data with filtering of measurement noises for predictive lithological mapping. Nat. Resour. Res. 30, 1033-1052.

[21] Guidici D., Clark M.L., 2017. One-Dimensional convolutional neural network land-cover classification of multi-seasonal hyperspectral imagery in the San Francisco Bay Area, California. Remote Sens. 9, 629.

[22] Guillot S., Le Fort P., 1995. Geochemical constraints on the bimodal origin of High Himalayan leucogranites. Lithos 35, 221-234.

[23] Gulli A., Pal S., 2017. Deep learning with Keras. Packt Publishing Ltd.

[24] Guo J., Li W., Li G., Jiao Y., Liang S., 2019. Application of Multi-Scale Integrated Geophysical Method in Prospecting Prediction of Zhaxikang Pb-Zn-Sb-Au Polymetallic Deposit. Earth Sci. 44, 2129-2142(In Chinese with English abstract).

[25] He C., Qin K., Zhao J., Evans N., Li J., Zhou Q. Li G., 2022. Multiple skarn generations related to composite leucogranites in the Cuonadong Sn-W-Be deposit, Himalaya. Ore Geol. Rev., 105161.

[26] Hochreiter S., Schmidhuber J., 1997. Long short-term memory. Neural comput. 9, 1735-1780.

[27] Hossin M., Sulaiman M.N., 2015. A review on evaluation metrics for data classification evaluations. Int.J.Data Mining Knowledge Manag.ement Process 5, 1-11.

[28] Hou Z., Cook N.J., 2009. Metallogenesis of the Tibetan collisional orogen: A review and introduction to the special issue. Ore Geol. Rev. 36, 2-24.

[29] Hu W., Li H., Pan L., Li W., Tao R., Du Q., 2020. Spatial-spectral feature extraction via deep ConvLSTM neural networks for hyperspectral image classification. IEEE Trans. Geosci. Remote Sens. 58, 4237-4250.

[30] Huang Y., Fu J., Li G., Zhang L., Liu H., 2019. Determination of Lalong Dome in South Tibet and New Discovery of Rare Metal Mineralization. Earth Sci. 44, 2197-2206(In Chinese with English abstract).

[31] Hunt G.R., 2017. Spectroscopic properties of rocks and minerals. Handbook of Physical Properties of Rocks 1, 295-385.

[32] Ioffe S., Szegedy C., 2015. Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shift. Proceedings of the 32nd International Conference on Machine Learning, PMLR 37, 448-456.

[33] Jiao Y., Huang X., Li G., Liang S., Guo J., 2019. Deep Structure and Mineralization of Zhaxikang Ore-Concentration Area, South Tibet: Evidence from Geophysics. Earth Sci. 44, 2117-2128(In Chinese with English abstract).

[34] Kiranyaz S., Avci O., Abdeljaber O., Ince T., Gabbouj M., Inman D.J., 2021. 1D convolutional neural networks and applications: A survey. Mechanical Systems and Signal Processing 151, 107398.

[35] Kiranyaz S., Ince T., Abdeljaber O., Avci O., Gabbouj M., 2019. 1-d convolutional neural networks for signal processing applications. IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP) 8360-8364.

[36] Krizhevsky A., Sutskever I., Hinton G.E., 2012. Imagenet classification with deep convolutional neural networks. Advances in Neural Information Processing Systems 25, 1097-1105.

[37] Le Fort P., Cuney M., Deniel C., France-Lanord C., Sheppard S., Upreti B.N., Vidal P., 1987. Crustal generation of the Himalayan leucogranites. Tectonophysics 134, 39-57.

[38] LeCun Y., Bengio Y., Hinton G., 2015. Deep learning. Nature 521, 436-444.

[39] Li G., Zhang L., Jiao Y., Xia X., Dong S., Fu J., Liang W., Zhang Z., Wu J., Dong L., 2017. First discovery and implications of Cuonadong superlarge Be-W-Sn polymetallic deposit in Himalayan metallogenic belt, southern Tibet. Mineral Deposits 36, 1003-1008(In Chinese with English abstract).

[40] Lipton Z.C., Berkowitz J., Elkan C., 2015. A critical review of recurrent neural networks for sequence learning. arXiv preprint arXiv:1506.00019.

[41] Liu C., Wang R., Linnen R. L., Wu F., Xie L., Liu X., 2023. Continuous Be mineralization from two-mica granite to pegmatite: Critical element enrichment processes in a Himalayan leucogranite pluton. Am. Mineral., 108(1), 31-41.

[42] Liu Z., Wu F., Liu X., Wang J., Yin R., Qiu Z., Ji W., Yang L., 2019. Mineralogical evidence for fractionation processes in the Himalayan leucogranites of the Ramba Dome, southern Tibet. Lithos 340, 71-86.

[43] Liu L., Zhu D., Wang Q., Cawood P.A., Stockli D.F., Stockli L.D., Lin C., Zhang J., Zhang L. Zhao Z., 2022. Leucogranite records multiple collisional orogenies. Geophys. Res. Lett. 49, e2021GL096817.

[44] Lui T.C., Gregory D.D., Anderson M., Lee W.S., Cowling S.A., 2022. Applying machine learning methods to predict geology using soil sample geochemistry. Applied Comput. Geosci. 16, 100094.

[45] Miclea A.V., Terebes R., Meza S., 2020. One dimensional convolutional neural networks and local binary patterns for hyperspectral image classification. IEEE International Conference on Automation, Quality and Testing, Robotics (AQTR) 1-6.

[46] Parsa M., Maghsoudi A., Yousefi M., Carranza E.J.M., 2017. Multifractal interpolation and spectrum-area fractal modeling of stream sediment geochemical data: Implications for mapping exploration targets. j.Afr. Earth Sci. 128, 5-15.

[47] Piippo S., Sadeghi M., Koivisto E., Skyttä P., Baker T., 2022. Semi-automated geological mapping and target generation from geochemical and magnetic data in Halkidiki region, Greece. Ore Geol. Rev. 142, 104714.

[48] Qin K., Zhao J., He C., Shi R., 2021. Discovery of the Qongjiagang giant lithium pegmatite deposit in Himalaya, Tibet, China. Acta Petrol. Sin. 37, 3277-86(In Chinese with English abstract).

[49] Reichstein M., Camps-Valls G., Stevens B., Jung M., Denzler J., Carvalhais N., 2019. Deep learning and process understanding for data-driven Earth system science. Nature 566, 195-204.

[50] Sherstinsky A., 2020. Fundamentals of recurrent neural network (RNN) and long short-term memory (LSTM) network. Physica D: Nonlinear Phenomena 404, 132306.

[51] Shirmard H., Farahbakhsh E., Muller D., Chandra R., 2022. A review of machine learning in processing remote sensing data for mineral exploration. Remote Sens. Environ. 268, 112750.

[52] Sun T., Li H., Wu K., Chen F., Zhu Z., Hu Z., 2020. Data-driven predictive modelling of mineral prospectivity using machine learning and deep learning methods: a case study from southern Jiangxi Province, China. Minerals 10, 102.

[53] Talebi H., Peeters L.J.M., Mueller U., Tolosana-Delgado R. van den Boogaart K.G., 2020. Towards geostatistical learning for the geosciences: A case study in improving the spatial awareness of spectral clustering. Math. Geosci. 52, 1035-1048.

[54] Van Houdt G., Mosquera C., Nápoles G., 2020. A review on the long short-term memory model. Artif. Intell. Rev. 53, 5929-5955.

[55] Wang Z., Zuo R., 2023. Mapping the spatial distribution of Himalayan leucogranites through fusion of multi-source geoscience data. Earth Sci. Front. 10.13745/j.esf.sf.2023.5.22 (In Chinese with English abstract).

[56] Wang Z., Zuo R., Dong Y., 2020. Mapping Himalayan leucogranites using a hybrid method of metric learning and support vector machine. Comput. Geosci. 138, 104455.

[57] Wang Z., Zuo R. Liu H., 2021a. Lithological mapping based on fully convolutional network and multi-source geological data. Remote Sens. 13, 4860.

[58] Wang Z., Zuo R., Jing L., 2021b. Fusion of Geochemical and Remote-Sensing Data forLithological Mapping Using Random Forest Metric Learning. Math. Geosci. 53, 1125-1145.

[59] Wang X., Zuo R., Wang Z., 2022. Lithological mapping using a convolutional neural network based on stream sediment geochemical survey data. Nat. Res. Res. 31(5), 2397-2412.

[60] Wang Z., Zuo R., 2022. Mineral prospectivity mapping using a joint singularity-based weighting method and long short-term memory network. Comput. Geosci.158, 104974.

[61] Wang Z., Zuo R., Yang F., 2023. Geological mapping using direct sampling and a convolutional neural network based on geochemical survey data. Math. Geosci. 10.1007/s11004-022-10023-z.

[62] Williams P.M., 2021. Statistical levelling of multi-element geochemical data. Applied Comput. Geosci. 10, 100060.

[63] Wu G., Chen G., Cheng Q., Zhang Z., Yang J., 2021. Unsupervised machine learning for lithological mapping using geochemical data in covered areas of Jining, China. Nat. Res. Res. 30, 1053-1068.

[64] Wu F., Liu Z., Liu X., Ji W., 2015. Himalayan leucogranite: Petrogenesis and implications to orogenesis and plateau uplift. Acta Petrol. Sin. 31, 1-36 (In Chinese with English abstract).

[65] Wu F., Liu X., Ji W., Wang J., Yang L., 2017. Highly fractionated granites: Recognition and research. Sci. China Earth Sci. 60, 1201-1219 (In Chinese with English abstract).

[66] Wu F., Liu X., Liu Z., Wang R., Xie L., Wang J., Ji W., Yang L., Liu C., Khanal G.P., 2020. Highly fractionated Himalayan leucogranites and associated rare-metal mineralization. Lithos 352, 105319.

[67] Wu H., Prasad S., 2017. Convolutional recurrent neural networks forhyperspectral data classification. Remote Sens. 9, 298.

[68] Xie L., Tao X., Wang R., Wu F., Liu C., Liu X., Li X., Zhang R., 2020. Highly fractionated leucogranites in the eastern Himalayan Cuonadong dome and related magmatic Be-Nb-Ta and hydrothermal Be-W-Sn mineralization. Lithos 354, 105286.

[69] Xie X., Yan M., Wang C., Li L., Shen H., 1989. Geochemical standard reference samples GSD 9-12, GSS 1-8 and GSR 1-6. Geostandards Newsletter 13, 83-179.

[70] Xie X., X., Mu X., Ren T., 1997. Geochemical mapping in China. j.Geochem. Explor. 60, 99-113.

[71] Xiong Y., Zuo R., 2020. Recognizing multivariate geochemical anomalies for mineral exploration by combining deep learning and one-class support vector machine. Comput. Geosci. 140, 104484.

[72] Xiong Y., Zuo R., Luo Z., Wang X., 2022. A Physically Constrained Variational Autoencoder for Geochemical Pattern Recognition. Math. Geosci. 54,783-806.

[73] Yang L., Liu X., Wang J., Wu F., 2019. Is Himalayan leucogranite a product by in situ partial melting of the Greater Himalayan Crystalline? A comparative study of leucosome and leucogranite from Nyalam, southern Tibet. Lithos 342, 542-556.

[74] Yin A., 2006. Cenozoic tectonic evolution of the Himalayan orogen as constrained by along-strike variation of structural geometry, exhumation history, and foreland sedimentation. Earth Sci. Rev. 76, 1-131.

[75] Yin B., Zuo R., Xiong Y., 2022. Mineral prospectivity mapping via gated recurrent unit model. Nat. Res. Res. 31, 2065-2079.

[76] Yu Y., Si X., Hu C., Zhang J., 2019. A review of recurrent neural networks: LSTM cells and network architectures. Neural comput. 31, 1235-1270.

[77] Zhang C., Zuo R., Xiong Y., 2021. Detection of the multivariate geochemical anomalies associated with mineralization using a deep convolutional neural network and a pixel-pair feature method. Applied Geochem. 130, 104994.

[78] Zuo R., Xu Y., 2023. Graph deep learning model for mapping mineral prospectivity. Math. Geosci. 55, 1-21. 10.1007/s11004-022-10015-z.

[79] Zuo R., Cheng Q., Agterberg F.P., 2009. Application of a hybrid method combining multilevel fuzzy comprehensive evaluation with asymmetric fuzzy relation analysis to mapping prospectivity. Ore Geol. Rev. 35, 101-108.

[80] Zuo R., Xiong Y., Wang Z., Wang J., Kreuzer O. P., 2023. A New generation of artificial intelligence algorithms for mineral prospectivity mapping. Nat. Res. Res. 32, 1859-1869.

[81] Zuo R., Wang Z., 2020. Effects of random negative training samples on mineral prospectivity mapping. Nat. Res. Res. 29, 3443-3455.

[82] Zuo R., Xiong Y., Wang J., Carranza E.J.M., 2019. Deep learning and its application in geochemical mapping. Earth-sci. rev. 192, 1-14.