| 潘北斗.不同类型矿床（点）空间关联关系定量研究—以中南半岛及临近地区为例[J].地质与勘探,2022,58(3):532-544 |
| 不同类型矿床（点）空间关联关系定量研究—以中南半岛及临近地区为例 |
| Quantitative study on spatial correlation of different types of deposits (occurrences)-Case study of Indochina Peninsula and adjacent areas |
| 投稿时间：2021-12-23 修订日期：2022-04-24 |
| DOI：10.12134/j.dzykt.2022.03.006 |
| 中文关键词: ROC曲线 定量 矿点 空间关联 中南半岛 |
| 英文关键词: ROC curve, mineral occurrence, quantitative analysis, spatial relationship, Central South Peninsula |
| 基金项目:中国冶金地质总局《大兴安岭成矿带成果及成矿预测研究》和西华师范大学博士科研启动项目（编号：18Q018）联
合资助 |
|
| 摘要点击次数: 1081 |
| 全文下载次数: 480 |
| 中文摘要: |
| 在具有较多已知矿床（点）的区域，从矿床（点）的空间分布模式中可以挖掘出成矿过程的重要信息，这些信息，对于认识区域成矿规律，指导区域找矿具有重要意义。不同矿床类型之间的空间关联，通常都是通过结合地质和成矿规律定性分析来研究，缺少定量、精确的研究方法。针对这个问题，本文提出了一种定量分析不同矿床（点）类型之间空间关联的方法—基于ROC曲线的矿床空间关系挖掘方法。该方法采用数据分析中常用的ROC曲线，通过构建矿床空间点缓冲区，利用GIS软件计算得到ROC曲线和曲线下面积AUC以及Youden指数等相关参数，利用ROC曲线和AUC等参数定量探讨不同矿床类型之间的空间关联。通过网络分析对不同矿床类型之间的空间关系定量指标进行可视化；通过社团检测算法找出众多不同矿床类型中存在较强关联关系的矿床类型群组。以中南半岛矿床点的空间分布为实例进行研究，挖掘出了该研究区9种主要矿床类型之间的空间关联规律。结果表明，研究区的热液型Pb矿、热液型Cu矿、热液型Pb-Zn矿、砂矿型Au矿四种主要矿床类型之间存在较强空间关联（AUC值0.74~0.88），在空间上相互靠近，相伴产出。热液型Sn矿、砂矿型Sn两类之间关联关系强烈（AUC值0.62~0.79）。热液型Au矿、热液型Fe矿、热液型Sb矿的空间分布模式较为独立，与其他矿床类型之间关联较弱。利用本文提出的方法可以定量挖掘出区域上不同矿床类型之间的空间关联关系，对于区域成矿研究和找矿实践能够提供重要信息。 |
| 英文摘要: |
| In the area with a number of known deposits, some key information of the ore-forming process can be extracted from the mineral deposit spatial distribution. These information is of great significance for understanding the regional metallogenic regularity and guiding regional mineral prospecting. Generally, the spatial relationship between different types of deposit is studied qualitatively by combining geological and metallogenic information, lacking of quantitative and accurate research methods. To tackle this problem, we proposed a novel method —the ROC curve based spatial relationship analysis model, to quantitatively analyze the spatial relationship between different types of mineral deposits. Our model is based on the ROC curve which is a commonly used tool in data analysis. The GIS software is used to construct the spatial buffer of the mineral deposits, and calculate the ROC curve, AUC (Area Under Curve) and Youden index. ROC curve and AUC are used to quantitatively explore the relationship between different deposit types. The quantitative index of spatial relationship between different deposit types is visualized by network analysis. The community detection algorithm is used to find the deposit type groups with strong correlation among many different deposit types. 9 main mineral deposit types in the Central-South Peninsula is used as a study case to validate our approach. The results show that there are strong spatial correlation (AUC 0.74 ~ 0.88) among the four main deposit types of hydrothermal Pb deposit, hydrothermal Cu deposit, hydrothermal Pb-Zn deposit and placer Au deposit in the study area, which are spatially close to each other and appear together. The correlation between hydrothermal Sn deposit and placer Sn deposit is strong (AUC value is 0.62 ~ 0.79). The spatial distribution patterns of hydrothermal Au deposit, hydrothermal Fe deposit and hydrothermal Sb deposit are relatively independent, and their correlation with other deposit types are weak. The method proposed in this paper can be used to quantitatively mine the spatial correlation between different deposit types in the region, which can provide important information for regional metallogenic research and prospecting practice. |
|
查看全文
查看/发表评论 下载PDF阅读器 |
| 关闭 |
