摘要
成矿活动是多方面地质要素的耦合产物,而矿床(或地质体)具有明显的群聚性分布特征,准确确定矿床分布与相关地质体空间关系成为研究矿床成因与成矿预测的重要环节。证据权法成矿预测模型通过引入成矿有利度(C),对这种空间关系进行了量化描述,并以单元格成矿概率值为基础进行区域成矿预测。在分形模型中,则采用分维数对矿床的分布群聚性进行了数值化定义。通过结合证据权法成矿预测模型与分形模型,一种新的确定成矿概率临界值的方法被提出并应用在宁南地区铅锌成矿预测中。通过对证据权法模型中成矿概率值进行分形分析,目标区单元格分维数值D=0.0010、0.0033,表现出目标区域成矿条件具有强烈群聚性分布特征;确定的成矿概率临界值(r=0.1184、0.0073)满足D.E.Boleneus等所提出的临界值确定原则;经与专家经验模式所确定的成矿概率临界值对比分析,表明该方法拟合的临界值结果在排除人为干扰因素的同时,具备更高的预测精度,在成矿预测中是可信的。
Metallogenic activity is the coupling result of many geological elements, while the deposit （or geological bodies） has obvious characteristics of clustering distribution, so the accurate determination of the spatial relationship between the distribution of ore deposits and the related geological bodies is a critical aspect of studying the ore genesis and metallogenic prediction. The metallogenic prediction model using weight - of - evidence method （WofE） gives a quantitative description of such spatial rela- tionship by introducing the concept of metallogenic favorability, and this model can also make regional metallogenic prediction on the basis of metallogenic probability of unit grids. In the fractal models, the clustering degree of ore deposits is given a value - oriented definition by fractal dimensions. A new method is proposed to determine the threshold of metallogenic probability by combining the WofE metallogenic prediction model with the fractal model, and this method is applied to predict Pb - Zn potentiality in Ningnan region, Sichuan province. By making a fractal analysis on the value of metallogenic probability in the WofE model, the fractal dimension D of the unit grids in the target area is equal to 0. 0010 and 0. 0033, showing that ore deposits in the target have an obvious characteristic of clustering distribution; the determined threshold point （r = 0.1184,0. 0073 ） meets the principle of determining marginal value, which was proposed by D. E. Boleneus et al. ; Compared with the threshold of metallogenic probability determined by the expertise model, the thresh- old fitted by this method has a higher forecast precision when man - made interference factors are excluded, and is credible in the metallogenic prediction.
出处
《矿业研究与开发》
CAS
北大核心
2012年第1期29-33,122,共6页
Mining Research and Development
基金
国家重点基础研究发展计划项目(2007CB411400)
关键词
证据权法
分形模型
成矿概率
多项式拟合
Weights of evidence method, Fractal model, Metal- logenic probability, Polynomial fitting
