【作者】 易桂花；

【导师】 倪师军；

【作者基本信息】 成都理工大学 ， 地球化学， 2011， 博士

【摘要】 本文以青海省东昆仑五龙沟金矿集中区为研究区,提出并建立了“基于单元格定量统计”的化探异常与遥感异常响应分析模型,建立了化探异常与遥感异常的响应模式；构建了“R-D成矿盲预测法”,依据化探异常与遥感异常的响应分析结果,遴选了地质异常因子,开展了成矿预测工作,取得了较好的效果。研究结果对于提高我国西部其它具有相似地质背景区域的找矿效率,加快资源勘查的进度具有重要的参考价值和应用前景。取得的主要研究成果如下：1、通过对研究区化探数据处理及评价,获得了化探异常变量。运用迭代法、趋势面法、分形法、聚类分析、因子分析等多种地球化学数据处理方法,通过整区与分区研究相结合,圈定出各类化探异常,揭示了研究区地球化学异常的空间展布规律。2、通过研究区遥感矿化蚀变信息提取和线性构造解译,获得了遥感异常变量。以典型蚀变矿物波谱特征为分析基础,改进了研究区ETM羟基异常和铁染异常的提取方法；利用ASTER遥感数据提取了研究区的含Al羟基、含Mg羟基、碳酸岩化和铁染异常。同时,初步提出了不同遥感蚀变异常的综合方法,综合结果为化探异常与遥感异常的响应分析提供了重要的遥感蚀变信息基础。在研究区线性构造解译基础上,对线性构造开展了密度、频度、异常方位等定量统计分析工作；通过研究区线性构造频度图和密度图,反映了研究区地质构造的空间展布特征。3、建立了研究区化探异常与遥感异常的响应分析模型。对化探异常与遥感异常的响应分析基础和优势进行了初步的探讨,建立了“基于单元格定量统计”的响应分析模型；提出了遥感异常与化探异常的“响应度”的概念。同时,基于GIS的空间分析功能,开展了化探异常与遥感异常的空间响应分析；在此基础上建立了化探异常与遥感异常的“强-强响应”、“强-弱响应”、“弱-弱响应”和“不响应”4种响应模式,讨论了化探异常与遥感异常不同响应模式的示矿效应。4、提出了“R-D成矿盲预测法”,并开展了相应的成矿预测工作。在成矿预测要素分析的基础上,利用化探异常与遥感异常的响应关系,遴选了9个地质异常变量,提出了“R-D成矿盲预测法”,克服了证据权法和特征分析法地质变量定性描述(存在与否或有利与否)的缺陷,将地质变量的空间特征作为预测因素参与到成矿预测中来；以1km×1km网格单元作为预测单元,将地质变量的异常等级、异常规模、异常展布方向等异常特征定量化,计算了预测单元的成矿有利度,预测了成矿远景区。通过与已知矿床(点)的对比分析,77%的已知金矿床(点)位于成矿远景区的影响范围内,取得了较好的效果。该方法较好地将“专家知识”与地质异常特征进行了有效地结合(或整合),计算简单、目的明确,是一种适合地质工作程度低、具研究区相同地质、地貌景观区的有效的靶区定位预测方法。更多还原

【Abstract】 Selecting Wu Long Gou gold minerals concentrated district in East Kunlun Mountain in Qinghai Province as study area, this paper put forward a corresponding analysis model of remote sensing and goechemical anomaly, called "unit-based quantitative statistic method", set up the pattern of "corresponding degrees" that means correspondence between geochemical anomaly and remote sensing anomaly, and formed the method of "R-D metallogenic blind prognosis". According to the corresponding analysis, geological anomaly factors are selected to perform metallogenic predictation, and some innovative results have been achieved. The study results are expected to have great referential value and application prospects for elevating ore-hunting efficiency and speeding up mineral resources prospecting in similar areas.The main results of the research are as follows:1. Geochemical anomaly variables are acquired by geochemical data processing and evaluation in study area. By using methods of iteration, tendency analysis, fractal method, cluster analysis, principal factor analysis, together with geochemical zonation, spatial distribution of geochemical anomalies is revealed and delimited in this area.2. Remote sensing anomaly variables are acquired by extracting mineral alteration information and explaining linear structures on remote sensing data getting from the study area. Based on spectral characteristics of typical altered mineral, the method is improved to extract ETM remote sensing data of hydroxyl and ferric anomaly. Information of A1-OH hydroxy anomaly, Mg-OH hydroxy alterations, carbonatization alterations and ferric alterations are extracted by applying ASTER remote sensing data. The integrated approach to different alterations anomalies reflected by remote sensing is preliminaryly presented. Results offer an important information basis for the corresponding analysis on geochemical anomaly and remote sensing anomaly. Based on the deciphering of the linear structures in studied area, a quantitative analysis is made on density, frequentness and abnormal position of the linear structure. Finally, the spatial characters of linear structures in studied area are reflected clearly through studying density and frequency diagram of the structure.3. The corresponding analysis model of remote sensing and geochemical anomaly is set up. By exploring corresponding analysis basis and advantages on geochemical anomaly and remote sensing anomaly, the model of "unit quantitative statistic method" is built, and the concept of "corresponding degrees" from correspondence between geochemical anomaly and remote sensing anomaly is put forward. Based on the spatial analysis of GIS, spatial corresponding analysis is made between geochemical anomaly and remote sensing anomaly. Then, the following 4 corresponding models of "intense-intense corresponding", "intense-weak corresponding", "weak-weak corresponding", and "no-corresponding" are put forward in the paper. Finally, the illumination of metallogenic prognosis in different corresponding models is explored in the study.4. The method of "R-D metallogenic blind prognosis" is formed. Based on analyzing the metallogenic prediction elements by using corresponding analysis of geochemical and remote sensing anomalies, 9 geological anomaly variables are selected and "R-D metallogenic blind prognosis" is put forward. These methods help to overcome defects in qualitative description of geological variables, in the weights-of-evidence analysis and characteristic analysis. Spatial characteristics of the geological variables are taken as prognosis element in metallogenic prognosis prediction. Taking 1km×1km grid cell as prognosis unit, anomaly grade, anomaly size, anomaly azimuth of geology variable are quantified in the study, the advantageous degree of metallogenic prognosis of prognosis unit is calculated and the metallogenic prospect area is hopefully predicted.Through comparative analysis between metallogenic prospect area and known deposits,77% of known gold deposits are located in the metallogenic prospect area. The "R-D metallogenic blind prognosis" method combines "expert knowledge" with geology anomaly characters (concordance) preferably and effectively. With the advantages of simple calculation and definite purpose, it is also an effective prediction method for target location, suitable for the areas of inadequate geological exploration, and of similar geology and landform features.更多还原