中国科技核心期刊
美国化学文摘社（CAS）数据库
美国EBSCO学术数据库
日本科学技术振兴机构数据库(JST)
夏甸金矿床金红石机器学习成因判别与成矿作用分析
- 作者：
王荣超|杨晓奇|杨晓鹏|高腾|唐伟洋|陈树杰|陈煜东
- 作者单位：
- （1. 招金矿业股份有限公司；2. 中南大学地球科学与信息物理学院）
基金项目：
- 国家自然科学基金项目（42202332）
详细信息：
- 作者简介：
- 王荣超（1972—），男，高级工程师，博士，研究方向为成矿理论与找矿勘查；E⁃mail：rongchaowang@126. com
- 通讯作者：
- 杨晓奇（1986—），男，高级工程师，从事矿产资源预测评价与勘查设计工作；E⁃mail：yangxiaoqi1986@126. com
- PDF下载
Genetic discrimination and mineralization analysis of rutile in the Xiadian Gold Deposit using machine learning
- English Author：
Wang Rongchao¹, Yang Xiaoqi¹, Yang Xiaopeng¹, Gao Teng¹, Tang Weiyang¹, Chen Shujie¹, Chen Yudong²
- Unit：
- （1. Zhaojin Mining Industry Co., Ltd.; 2. School of Geosciences and Info‑physics, Central South University）
- 摘要
- 在线预览
- 参考文献
摘要:
金红石是胶东金矿床中的常见副矿物，但金红石成因及蕴含的成矿信息仍不明确。对招平断裂南端的夏甸金矿床金红石展开了岩相学、原位地球化学分析、机器学习等综合研究，结果显示，夏甸金矿床金红石主要发育于招平断裂下盘的硅化、黄铁绢英岩化蚀变岩中，呈碎裂-交代结构，与黄铁矿、黄铜矿等矿石矿物共生。通过系统收集全球已发表不同成因类型及矿床类型中金红石数据，采用随机森林模型，针对夏甸金矿床金红石微量元素特征开展成因判别，发现夏甸金矿床金红石具有典型热液成因特征，能够指示金成矿作用。此外，利用随机森林模型将金红石判别为“胶东型”金矿床成因类型，其较高的真正类率指示该类矿床与其他类型矿床金红石微量元素特征存在显著区别。其中，对判别起重要作用的元素为 Cr、Si、W、Fe、Cu、V、Ta。这些元素富集指示该区基底的高级变质岩在成矿作用期间发生元素活化，并进入金成矿系统，同时指示强烈的水岩反应在夏甸金矿床的形成过程中作用显著。
关键词:
金红石；元素地球化学；机器学习；成因判别；夏甸金矿床；招平断裂
Abstract:
Rutile is a common accessory mineral in the Jiaodong gold deposits, yet its genetie significance and metallogenic implications remain poorly constrained. This study integrates petrography, in-situ geochemical analysis, and machine learning to investigate rutile from the Xiadian Gold Deposit at the southern end of the Zhaoping fault. Results reveal that rutile of the Xiadian Gold Deposit primarily occurs in silicified and sericitized alteration rocks beneath the Zhaoping fault, exhibiting cataclastic-metasomatic textures and coexisting with ore minerals such as pyriteand chalcopyrite. By compiling a global dataset of rutile from perse genetic and deposit types, a random forest model was applied to discriminate the origin of rutile based on its trace element characteristics in the Xiadian Gold Deposit. The model identifies the rutile of the Xiadian Gold Deposit as typical hydrothermal in origin, closely linked to gold mineralization. Furthermore, the rutile is classified as "Jiaodong-type"gold deposit genesis with high true-class rates indicating distinct trace element signatures compared to other deposit types. Among them, Cr, Si, W, Fe, Cu, V, and Ta are the elements that play important roles in discrimination. Enrichment of these elements suggests elemental mobilization from high-grade metamorphic basement rocks during mineralization and highlights the critical role of intense water-rock interactions in the formation of the Xiadian Gold Deposit.
Keywords:
rutile; elemental geochemistry; machine learning; genetic discrimination; Xiadian Gold Deposit; Zhao⁃ ping fault
