基于数值模拟与勘查数据协同驱动的矿产定量预测——以凡口铅锌矿为例  

Numerical Modeling and Exploration Data Coupled-driven Mineral Prospectivity Mapping:A Case Study of Fankou Pb-Zn Deposit

作　　者：肖凡 陈信宇[1] XIAO Fan;CHEN Xinyu(School of Earth Sciences and Engineering,Sun Yat-Sen University,Zhuhai 519000,Guangdong,China;Guangdong Key Laboratory of Geological Process and Mineral Resources Exploration,Zhuhai 519000,Guangdong,China;Southern Laboratory of Ocean Science and Engineering,Zhuhai 519000,Guangdong,China)

机构地区：[1]中山大学地球科学与工程学院,广东珠海519000 [2]广东省地质过程与矿产资源探查重点实验室,广东珠海519000 [3]南方海洋科学与工程广东省实验室(珠海),广东珠海519000

出　　处：《大地构造与成矿学》2025年第2期298-316,共19页Geotectonica et Metallogenia

基　　金：国家重点研发计划项目(2022YFF0801201);广东省引进人才创新创业团队项目(2021ZT09H399)联合资助。

摘　　要：之前的成矿预测主要是基于描述性矿床模型,即勘查模型,对地、物、化、遥等各种找矿异常信息进行识别提取与关联融合,评估成矿有利度,从而圈定可能成矿的潜力区。然而,描述性矿床模型实质上是对复杂成矿系统的一种经验性的简化表示,主要强调矿产勘查中可直接观测的显式成矿要素,即找矿标志或找矿异常信息,而诸如奇异应力应变、含矿流体运聚、金属物质沉淀等有关成矿过程的物理‒化学参量的隐式成矿要素,由于难以被直接观测,往往在描述性矿床模型中被忽略。成矿预测模型中这些示踪成矿事件的隐式“指纹”信息的缺失,从根本上导致了成矿预测结果的精度低、不确定性增强。为此,本文提出了使用多场耦合数值模拟方法刻画和表征成矿过程中金属富集的有关物理‒化学参量的隐式成矿要素,继而与表示勘查标志异常信息的显式成矿要素进行融合,构建数值模拟与勘查数据协同驱动的成矿预测模型。以凡口铅锌矿床为例,构建力‒热‒流‒化学四场耦合的成矿动力学数值模拟模型,在时空尺度上再现成矿期流体流动和金属沉淀过程,从而获得应力应变、温度变化、流体通量等表征隐式成矿要素的物理‒化学参量。将这些参量作为预测变量,融合到成矿预测模型中,即以应力、应变、温度、流体通量、成矿元素浓度等物理‒化学模拟参量以及有利地层、有利构造等空间变量作为成矿预测模型的特征变量,利用随机森林机器学习算法进行信息融合,计算成矿潜力,对研究区深边部找矿前景进行了预测评价,并与基于描述性矿床模型的数据驱动预测方法结果进行对比。最终得出以下结论:①建立复杂成矿系统的数学‒物理表征模型,利用数值模拟方法表征与成矿作用具有成因联系的隐式成矿条件参数,是成矿预测模型中融合矿床成因模型的有效Traditional mineral prospectivity mapping relies predominantly on descriptive deposit models,commonly referred to as exploration models.These models have guided the identification,extraction,and integration of geoinformation,including geological,geophysical,geochemical,and remote sensing anomalies,as well as the selection of favorable exploration targets.However,descriptive deposit models are inherently limited by the empirical simplification of complex mineral systems.They primarily focus on explicit features that are observable in mineralizing systems,such as mineral indicators or exploration anomalies,while implicit features related to physical-chemical conditions of mineralization processes,such as stress-strain anomalies and mineralizing fluid migration,are typically difficult to observe directly,leading to their frequent omission by descriptive deposit models.This fundamental limitation has posed significant challenges in terms of achieving high accuracy and reducing uncertainty in mineral prospectivity mapping.To address this critical issue,this study proposes a numerical modeling-exploration data-coupled mineral prospectivity mapping method.It utilizes physical-chemical parameters derived by the multi-field coupled numerical modeling method to characterize and represent the implicit features of a mineral system.These implicit features are then integrated with explicit features represented by anomalies identified from exploration data to predict mineral prospectivity.As a case study,the Fankou Pb-Zn deposit was used to illustrate the proposed mineral prospectivity modeling method.The four-field coupled numerical simulation model incorporating thermal,hydrological,chemical,and mechanical processes was constructed to reproduce deformation,fluid flow,and metal precipitation results during the mineralization period at both spatial and temporal scales.This model enables the characterization of physical-chemical parameters representing implicit mineralization features,such as stress-strain anomalies,temperature

关 键 词：数值模拟 随机森林 空间分析 成矿预测 凡口铅锌矿 

分 类 号：P612[天文地球—矿床学]