| 基于改良信息量法的萤石成矿“甜点区”预测 |
| 点此下载全文 |
| 引用本文：黄超,蓝升,杨相,黄声济,温培焱,罗林山,刘凤吉,陈灵.2026.基于改良信息量法的萤石成矿“甜点区”预测[J].地球学报,47(2):409-420. |
| DOI：10.3975/cagsb.2025.101102 |
| 摘要点击次数: 712 |
| 全文下载次数: 53 |
|
| 基金项目:本文由赣州市社科规划课题(2025年)(编号: 2025-NDJJ19-1528)资助。 |
|
| 中文摘要:针对赣南萤石成矿预测区域性不足、空间分辨率低及主观性强等问题, 本文提出改良栅格信息量法。基于区域成矿规律, 筛选区域深大断裂、断层、地层、水系、地形地貌、F与CaO地球化学7类控矿因子; 创新性地将面状矿权数据栅格化处理, 量化因子信息量值, 并基于ArcGIS平台构建多因子叠加预测模型; 利用受试者特征(ROC)曲线分析与补充调查验证预测结果。结果表明: 地层中燕山期黑云母花岗岩信息量值最高(1.520); 断层、水系距离与信息量值呈负相关, F地球化学异常与成矿显著正相关; 距离区域深大断裂8~10 km带出现信息量峰值, CaO因子受背景干扰失效。预测圈定Ⅰ级、Ⅱ级成矿甜点区, 分别占研究区面积5.27%和14.87%, 划分出7条重点成矿带。模型验证曲线下方面积(AUC)值达0.818, 新发现矿点83.3%位于甜点区内, 表明预测结果准确。改良方法显著提升了预测空间分辨率, 成果符合地质实际, 可为赣南萤石矿权设置与勘查部署提供科学依据。 |
| 中文关键词:赣南地区 萤石 矿床地质特征 GIS空间分析 信息量法 成矿预测 |
| |
| Prediction of Fluorite Mineralization “Sweet Spot” Based on an Improved Information Method |
|
|
| Abstract:In this study, we propose an improved raster-based information method to address the issues of insufficient regionality, low spatial resolution, and high subjectivity in fluorite metallogenic prediction in southern Gannan. Based on the regional metallogenic law, seven types of ore-controlling factors were screened: regional deep-seated faults, faults, stratigraphy, hydrographic networks, topography and geomorphology, and geochemical anomalies of F and CaO. The surface element data were innovatively converted into raster format, and the factor information values were quantified. A multi-factor overlay prediction model was constructed using ArcGIS, and the prediction results were validated using Receiver Operating Characteristic (ROC) curve analysis and supplemental field surveys. Results indicate that the information value of the Yanshanian biotite granite is the highest (1.520). The distance to faults and hydrographic networks shows a negative correlation with the information value, whereas the F geochemical anomalies are significantly and positively correlated with mineralization. The peak information value occurs within the 8–10 km range from deep-seated faults. The CaO factor was deemed ineffective owing to background interference. The prediction delineates Grade I and Grade II sweet spots, accounting for 5.27% and 14.87% of the study area, respectively, and identifies seven key metallogenic belts. The area under the ROC curve is 0.818, and 83.3% of the newly discovered mineral occurrences fall within the predicted target zones, demonstrating the accuracy of the prediction. The improved method significantly enhances the spatial resolution of metallogenic prediction, and the results align well with the geological reality, providing a scientific basis for the planning and exploration allocation of fluorite mining rights in southern Gannan. |
| keywords:Gannan region fluorite geological characteristics of the deposit GIS spatial analysis information method metallogenic prediction |
| 查看全文 查看/发表评论 下载PDF阅读器 |
|
|
|
|
|
