【作者】 左仁广；

【导师】 赵鹏大； 成秋明； F.P.Agterberg；

【作者基本信息】 中国地质大学 ， 矿产普查与勘探， 2009， 博士

【摘要】 当今找矿难度日益增大，如何提高找矿效果一直是国内外地质学者关注的焦点问题之一。阻碍当前矿产勘查取得突破性进展的主要原因之一是成矿理论和勘查技术方法的滞后。因而，必须加强成矿地质理论探索和实验研究，深入了解各类矿床形成的环境和条件以及矿床分布规律与产出特征；同时，加强找矿技术方法研究，进一步查明指示矿床存在的各种标志和现象，有针对性地开发识别、获取、加工、分析和解释海量找矿信息的手段、工具和软件，研究深层次、隐性及难以识别的找矿信息的识别及提取方法。这不仅仅是一个技术问题，其中还隐含了基础理论问题，如支持弱信息识别、复合信息及混杂信息分离的基础理论等，而这些问题无疑正是制约当今找矿突破的世界性难题。赵鹏大等(1991)提出的以“求异”思维为基础的“地质异常”成矿预测理论突破了矿产勘查中“相似类比”思维的约束，并经过十余年的不断完善和发展，逐渐形成了以“地质异常”、“成矿多样性”和“矿床谱系”为核心的“三联式”定量成矿预测理论与方法，业已应用于固体矿产预测评价和油气资源勘探开发，并在山东、云南等地的找矿实践中取得了成效。进入21世纪，矿产资源勘查面临新形势：传统的矿产资源短缺，尤其是战略性、支柱性矿产，如石油、富铁、锰、铬、钾盐等形势更为严峻，出现了一批传统资源面临枯竭的“危机矿山和矿业城市”；未发现矿产很多属于难识别、难发现和难利用的复杂矿床(赵鹏大等，2002)。当前，主要以找深部矿和难识别矿为主，尤其在勘探程度较高的中、东部地区。因此，如何识别、提取和圈定新型的、隐型(式)的和深层次的成矿地质信息成为当前矿产预测评价的关键。在这种形式下，研发在“三联式”成矿预测理论与方法指导下的能有效识别、提取和圈定地质异常的专业GIS软件就显得格外重要。此外，由于地质本身的变化性和复杂性，如矿床类型的多样性，矿床成因的复杂性，控矿因素的隐蔽性和找矿信息的多解性(赵鹏大，2007)以及人类认识的不完备性等因素，导致了矿产预测结果具有不确定性并常常因人而异。矿产定量预测与评价是在不确定性下制定最优决策的工作，是在各种可能的决策和所对应的可能结果(或称“决策谱”)中选择一个最佳结果，即在不漏失或最少漏失成矿远景区或矿体的前提下最大限度的缩小需要进行详细工作的地区范围，达到成功和收益最大，损失和消耗最小(赵鹏大，2007)。因此，矿产定量预测与评价的主要目的之一是查明矿产预测不确定性的来源，寻求有效的方法与途径减少不确定性，使其矿产预测的正面结果概率最大，负面结果概率最小。20世纪90年代以来，世界各地的地质学家已开始注意并致力于研究矿产预测与评价的不确定性问题，该项研究业已成为探索“热点”。本文以国土资源大调查重大项目“全国矿产资源潜力评价”下属“矿产资源定量化预测新方法研究”为依托，在“三联式”定量成矿预测理论与方法的指导下，紧密围绕地质异常理论和不确定性评价这个主题，开展了地质异常的地质基础和数学基础，地质异常识别、提取和圈定技术与方法，矿产预测不确定性评价的途径与方法等研究，研发了基于地质异常的矿产预测与评价GIS软件，并以西藏冈底斯斑岩型铜矿为例，圈定了斑岩铜矿找矿远景区并评价了其不确定性。本研究主要包括以下一些内容。①探讨了地质异常的地质基础和数学基础地质异常是地质异常事件作用的结果，而反过来，地质异常又是各种事件的策源地和诱因，因此，可认为地质异常的地质基础为地质异常事件。极值和异常值相对地质背景来说都可认为是地质异常，地质异常的观测值位于分布的尾部，而极值分析正是研究在超常大(或小)水平上量化过程的随机性状，并估计比任何已观测水平更为极端事件的概率，因此，地质异常属于极值理论研究的范畴。②搭建了矿产预测不确定性评价主流程基于GIS矿产预测不确定性分为矿产定位预测不确定性(包括数据的不确定性、预测模型与预测方法的不确定性、矿床产出空间位置的不确定性)和资源潜力预测不确定性(包括未发现矿床(点)数的不确定性，未发现矿床(点)品位、吨位及资源量的不确定性)，总结每种类型的不确定性主要来源和研究内容；并用模糊集评价矿产预测中的不确定性，引入了隶属度的概念，从量上把握和处理矿产预测中复杂性和模糊不确定性等问题，可反映矿产预测结果的可靠性、可能性、误差和风险的大小。研究了矿产预测不确定性表达模型和传播模型及相应的计算方法，提出了降低矿产预测评价不确定性的途径。从而初步搭建了矿产预测不确定性评价体系，该体系包括矿产预测不确定性来源、分类→矿产预测不确定性评价→矿产预测不确定性表达与传播→降低矿产预测不确定性途径与方法。③探讨了矿产预测中定性数据的不确定性评价方法以“全国重要成矿区带数据库”为例，对各类型数据的字段数和存储空间进行了统计分析，研究发现定性数据在字段总数和存储空间都占主导地位，可见研究定性数据的不确定性是矿产预测不确定性评价中关键，并把定性数据分为文字描述型、分类码和顺序码等3类，统计分析这3类数据的字段数和存储空间，发现文字描述型数据和分类码数据是定性不确定性评价的重点。运用语言学算子对描述型数据进行不确定性评价，采用定性排序和定量转化方法评价分类码和顺序码数据不确定性，并以地层和断裂方位为例评价了定性数据的不确定性。④发展了地质异常识别、提取与圈定技术并研发基于地质异常的矿产预测与评价GIS系统基于GIS开发了极值理论、模糊数学及C-A法等圈定地质异常方法，发展了地质异常识别、提取与圈定技术，并研发了基于地质异常矿产预测评价GIS软件。该系统基本实现了矿产预测评价主流程的信息化与定量化，它具有地质异常识别与提取、变量变换、变量构置与优化、异常圈定、预测远景区圈定与优选等功能，实现了对矿产定位预测与矿产潜力预测的不确定性评价。这些对提取与圈定新型的、隐型(式)的和深层次的成矿地质信息和评价矿产预测的风险有一定的帮助。⑤圈定了西藏冈底斯斑岩型铜矿找矿远景区并评价其不确定性收集西藏冈底斯东段1：50万地理、地质、矿产地、地球化学、遥感、航磁和重力数据，建立了矿产预测基础空间数据库。在此基础上，利用基于地质异常的矿产预测评价系统对冈底斯斑岩型铜矿进行地质异常信息识别、提取与圈定；利用非线性方法-奇异性指数绘制法进行了Cu等异常信息提取；利用不对称模糊关系分析计算了预测变量的权重；利用多层模糊综合预测了斑岩型铜的找矿远景区，为在该区进行进一步矿产资源勘查与评价提供了参考依据。研究结果表明(1)奇异性指数能有效提取不同地质背景下的弱缓异常；(2)模糊不对称关系考虑了变量间的对称与不对称关系，其计算的预测变量的权重更能反映地质变量的关系，减少了预测变量权重的不确定性；(3)多层次模糊关系可有效的综合多层次多种类预测变量，并可评价预测结果的不确定性。综上所述，对基于地质异常的矿产预测与不确定性评价的理论、方法与技术做了有益探索。在地质异常理论与方法方面，研究了地质异常的地质与数学基础，发展了地质异常识别、提取与圈定技术，实现了基于地质异常的矿产预测与评价系统；在矿产预测不确定性评价方面，搭建了矿产预测不确定性评价的主流程，探讨了评价矿产预测不确定性的方法与技术，并圈定了西藏冈底斯斑岩型铜矿的找矿远景区并评价了其模糊不确定性，检验了基于地质异常的矿产预测与评价系统的实用性及矿产预测不确定性评价主流程的可操作性。上述工作的开展，为更广泛的应用地质异常理论开展矿产预测及不确定性评价提供理论、方法与软件支持，以期对提取与圈定新型的、隐型(式)的和深层次的成矿地质信息和减小矿产预测不确定性有一定的帮助。更多还原

【Abstract】 Given the increasing difficulties in exploring for mineral resources, improving exploration efficiency has become a major objective. The traditional exploration theories and methods are useful but also a major factor impeding mineral resources exploration. Therefore, we need to develop new geological exploration theory and methods to gain a better understanding of the processes that control ore body formation and mineralization, spatial and temporal rules of deposit distributions, and the economic properties of deposits. We also need to improve prospecting techniques so as to identify various indicators of mineralization, as well as software packages that can effectively identify, manipulate, analyze, and interpret massive mineralization information. Identification of deep and hidden information is not only a technical issue, but also an important scientific concern, such as the fundamental theory, identification of weak and complex information, and separation of hybrid data. Undoubtedly, these problems are worldwide and constrain the development of today’s breakthroughs in mineral resources exploration. The geological anomaly theory proposed by Zhao (1991) breaks through the bound of the "similar analogy" theory, and has gradually led to a new theory and method of "Three Components" , which consists of geological anomaly, diversity of mineralization, and the spectrum of mineral deposits (Zhao et al., 2001) . The theory and method have been widely used for predicting solid mineral resources and oil and gas exploration, and have gained wide application to information from ore fields in Shandong and Yunnan Province, China.In recent years, mineral resources prospecting has encountered two new situations: (1) There is a shortage of traditional mineral resources, especially for strategic and support commodities such as oil, iron, manganese, chromium, and potassium, resulting in a group of crisis mines and crisis mine cities; and (2) many undiscovered mineral deposits are difficult to identify, detect, and use. At present, we focus mainly on finding the deep and difficult-to-identify mineral deposits, especially in better explored areas such as in mid-China and eastern China. The identification, extraction and delineation of new, hidden, and deep ore-forming information play an important role in the prediction and assessment of mineral resources. It is important therefore to develop an effective tool to identify, extract, and delineate geological anomalies based on GIS under the guidance of "Three Components Prediction Theory and Methods". In addition, mineral resources prediction and assessment entail high risk and uncertainty because of the complexity and variability of geological objects, such as the diversity of mineral deposit types, the complexity of mineral deposit genesis, the implicitness of mineral deposit controlling factors, and the non-unique understanding of exploration information and because most mineral deposits are located subsurface at variable depths. How to identify the sources of uncertainty and how to improve the efficiency of exploration is not only a goal of all geologists but also a major scientific issue. Only through a comprehensive analysis of the sources of uncertainty in mineral resources prediction and through the quantitative evaluation of uncertainty can we find the approach and method to reduce risks and increase efficiency in exploration.This research is supported by national mineral resources potential prediction and assessment. Using "Three Components Prediction Theory and Methods" and geological anomaly as a guide, I conduct research on the geological and mathematical foundation of geological anomaly, the technology and method to identify, extract, and delineate the new, hidden, and deep ore-forming information, and the uncertainty in mineral resources prediction and assessment; I also construct a mineral resources prediction and assessment model based on geological anomalies. For demonstration purposes, the Gangdese porphyry copper belt will be studied as an example.The following conclusions can be obtained from these studies:(1) Discussion of the geological and mathematical foundation of geological anomalyGeological anomaly is the product of the evolution of and the interaction between processes affecting the Earth’s geological layers during different geological periods. The features of any geological anomaly and the type and size of mineral resources are determined by rock age, tectonic setting, geological environment, and rock type. With the evolution of geological history, the early formation of the geological anomalies will evolve. Therefore, the geological anomalies have an evolutionary sequence in space and time. All the geological characteristics related to mineralization, including the conditions of mineralization and spatial and temporal ore-controlling factors, are represented as the geological anomaly events in the process of geological evolution. Therefore, the foundation of geological anomaly is the geological event, and the geological anomaly is the result of the succession of geological events.Extreme value and anomaly value in geology can be regarded as geological anomalies, and their values often occur in the tail of the frequency distribution. Extreme analysis is used to quantify random characters at either ultra-large or small scale, and to estimate the probability of extreme events. Therefore, geological anomaly falls within the scope of extreme value theory.(2) Research on the method to identify, extract, and delineate geological anomaly, and development of the geoanomaly-based mineral resources prediction and assessment systemSome indicators of ore, such as space morphology and spatial configuration features, are presented to identify geological anomalies; some methods, including extreme value theory (EVT, Fuzzy mathematic, and concentration-area(C-A), are used to delineate the geological anomalies; and some methods, such as evidence of weights and fuzzy logic, are used to integrate multi-variables. Then the GIS-based mineral resources prediction and evaluation are constructed. The system, comprising eight modules, for geological anomaly identification and extraction, variables transformation, mineral resources prediction and assessment, evaluation of uncertainty, etc., has paved the main information course in mineral resources prediction and assessment. These are undoubtedly helpful to extract and delineate the new, hidden, and deep-mineralization information and evaluate the uncertainty.(3) Set up the main flow of uncertainty evaluation in the mineral resources prospectingThe main sources of uncertainty arise from two factors. The first is known as the geological uncertainty, including the variability and complexity of natural phenomena. The second known as the process of mineral resources prediction and assessment. The uncertainty is transformed from the previous stage to the next stage, resulting in a considerable amount of uncertainty accumulation and dissemination. The uncertainty can be classified as the uncertainty in mineral resources location prediction and the uncertainty in mineral resources potential prediction. Both of the uncertainties have been categorized into the uncertainty of spatial data, the uncertainty of prediction model, the uncertainty of undiscovered deposits, and the uncertainty of grade and tonnage. All these uncertainties are introduced in details and evaluated by fuzzy sets, where the fuzzy numbers are used to express the reliability, probability, and variance of the results. Then the expression and propagation model of uncertainty, and some methods to reduce the uncertainty in mineral resources exploration, are proposed. Based on previous research, the main flow of the uncertainty evaluation in the mineral resources prediction was set up, including the sources and classification of uncertainty, evaluation of uncertainty, the expression and propagation of uncertainty, and how to reduce the uncertainty.(4) Preliminary study of uncertainty evaluation in geological qualitative dataTaking the "National mineral resources database" as an example, the types of data fields and storage space are calculated. The results show that the qualitative data are still dominant in the massive data. Therefore, how to evaluate uncertainty in qualitative data is vital for mineral resources exploration. The qualitative data can be classified into two types. The first relates to description type, which is evaluated by using operators of the linguistic variables. The second is codes type, which is evaluated by qualitative sorting and quantitative transformation. Two examples, permissive strata and faults, demonstrate these methods.(5) Mineral resources prospecting, and evaluating its uncertainty for Gangdese porphyry copper depositsBasic spatial databases, including geology database, ore deposit database, geophysics database, geochemistry database, and remote sensing database, were established at the scale of 1:500,000, and the geoanomaly-based mineral resources prediction and assessment system was used to identify and extract geological anomalies, and to integrate multi-geovariables and evaluate the uncertainty. The results demonstrate that (1) the mapping singularity technique is a useful tool to separate weak anomalies from complex background; (2) asymmetric fuzzy association analysis can uncover both direct and indirect relations between variables, which generally more closely reflects the real relationships between geosciences variables, and can lead to better results; (3) the multilevel fuzzy comprehensive evaluation can efficiently efficiently integrate multilevel and multi-sources variables and handle uncertainty due to vagueness of classification in mineral prospectivity mapping.In a word, this dissertation mainly focuses on the theroy and method of both geoanomaly-based and uncertainty evaluation in mineral resources prediction and assessment. For the theory and method of geological anomaly, this paper (1) discusses the geological and mathematical foundation of geological anomaly; (2) develope of the methods to identify, extract, and delineate the new, hidden, and deep ore-forming information; and (3) develop geoanomaly-based mineral resources prospecting system. For the theory and method of uncertainty evaluation, this paper (1) sets up the main information flow for the evaluation of uncertainty in mineral resources prospecting and (2) discusses how to evaluate the uncertainty. The Gangdese porphyry copper belt in Tibet is then chosen as a study area to demonstrate that the geoanomaly-based mineral resources prospecting system is a practical and operational tool to map prospectivity and evaluate its uncertainty.更多还原