Factor analysis applied to regional geochemical data: problems and possibilities

A large regional geochemical data set of C-horizon podzol samples from a 188,000 km² area in the European Arctic, analysed for more than 50 elements, was used to test the influence of different variants of factor analysis on the results extracted. Due to the nature of regional geochemical data (neither normal nor log-normal, strongly skewed, often multi-modal data distributions), the simplest methods of factor analysis with the least statistical assumptions perform best. As a result of this test it can generally be suggested to use principal factor analysis with an orthogonal rotation for such data. Selecting the number of factors to extract is difficult, however, the scree plot provides some useful help. For the test data, a low number of extracted factors gave the most informative results. Deleting or adding just 1 element in the input matrix can drastically change the results of factor analysis. Given that selection of elements is often rather based on availability of analytical packages (or detection limits) than on geochemical reasoning this is a disturbing result. Factor analysis revealed the most interesting data structures when a low number of variables were entered. A graphical presentation of the loadings and a simple, automated mapping technique allows extraction of the most interesting results of different factor analyses in one glance. Results presented here underline the importance of careful univariate data analysis prior to entering factor analysis. Outliers should be removed from the dataset and different populations present in the data should be treated separately. Factor analysis can be used to explore a large data set for hidden multivariate data structures.     

Variogram models for regional groundwater geochemical data

Four variogram models for regional groundwater geochemical data are presented. These models were developed from an empirical study of the sample variograms for more than 10 elements in groundwaters from two geologic regions in the Plainview quandrangle, Texas. A procedure is given for the estimation of the variogram in the isotropic and anisotropic case. The variograms were found useful for quantifying the differences in spatial variability for elements within a geologic unit and for elements in different geologic units. Additionally, the variogram analysis enables assessment of the assumption of statistical independence of regional samples which is commonly used in many statistical procedures. The estimated variograms are used in computation of kriged estimates for the Plainview quadrangle data. The results indicate that an inverse distance weighting model was superior for prediction than simple kriging with the particular variograms used.     

因子分析与地统计学在化探数据分析中的应用

以铜陵矿集区土壤勘查地球化学数据为实例,应用因子分析方法获取了地球化学数据中的多个主因子信息,并利用地统计学方法开展了各主因子的空间变异分析和插值研究。研究结果显示,因子分析得到的各主因子对应于不同的成矿信息,将因子分析与地统计学分析和插值方法相结合,可以更好的展现各主因子得分的空间分布趋势以及与已知成矿信息的关联程度,进而服务于成矿预测和找矿勘探工作。     

大数据科学从信息化、模式化到智能化:现代地球化学应用研究的新范式

地球科学领域处在信息科学与信息化社会时代,基于大数据战略驱动的现代地球科学正在蓬勃兴起.诞生于近代的地球化学具有天然的结构化信息科学属性.提出建立从信息化、模式化到智能化的地球化学大数据信息科学应用研究总体思路,即基于地球化学大数据首先建立信息化系统,运用地球系统科学方法理论建立成矿地球化学分带富集模式与生态地球化学累...     

Robust statistics and geochemical data analysis

Advantages of robust procedures over ordinary least-squares procedures in geochemical data analysis is demonstrated using NURE data from the Hot Springs Quadrangle, South Dakota, U.S.A. Robust principal components analysis with 5% multivariate trimming successfully guarded the analysis against perturbations by outliers and increased the number of interpretable factors. Regression with SINE estimates significantly increased the goodness-of-fit of the regression and improved the correspondence of delineated anomalies with known uranium prospects. Because of the ubiquitous existence of outliers in geochemical data, robust statistical procedures are suggested as routine procedures to replace ordinary least-squares procedures.     

因子分析在地球化学数据处理中的应用

因子分析是一种多元统计分析方法,以往地质研究中常用其解决复杂的地质成因及矿化叠加问题。本文利用R型因子分析对岷县寨上-马坞调查区的1:20万水系沉积物样品数据进行处理,计算因子得分,圈定因子得分异常图,并与区内的各种地质作用相联系,取得了一定的效果;利用R-Q型因子分析对寨上及附近的原生晕样品进行分析,通过对矿体及构造带样品的分析,可以划分矿化类型;通过对地层样品的分析,可以划分地层并得出各地层的特征组合元素。因此利用因子分析可以为地球化学数据解释提供重要信息。     

Semi-hierarchical correspondence cluster analysis and regional geochemical pattern recognition

Semi-hierarchical correspondence cluster analysis (SHCCA), firstly developed in this paper, extracts the main advantages of correspondence analysis, hierarchical and non-hierarchical cluster analysis, and unifies the R- and Q-mode cluster analysis of large data set. A systemic program to recognize the regional geochemical patterns is built up based on this method. With this program, the complex tasks for data interpretation can be achieved by simple processes, and important geochemical information can be displayed by a single diagram, i.e. the multivariate regional geochemical image. As one of the applied examples of this program, the regional geochemical pattern recognition for a shallow covered area around Tahe in Heilongjiang Province is introduced. The results show that many hidden geochemical patterns related to the lithologies, structures, ore-forming conditions and prospecting targets etc are revealed by the geochemical image, and that the main geochemical patterns are related with certain geological and gravitational patterns. By finding contrasts between geochemical patterns and geological or gravitational patterns, the SHCCA results assist the geological mapping in this area. Geochemical data obtained in Chinese regional geochemical exploration provides useful information regarding geology and minerals, and the method described in this paper provides a new way to examine this type of resource.     

R型聚类分析在区域化探元素分组中的作用探讨

在处理化探数据方面,R型聚类分析是应用较为广泛的一种数理统计方法。随着研究的深入发现,R型聚类分析的结果并非总是对元素分组起到积极有效的作用。这里对大石寨地区的二叠纪地层、白垩纪地层、黑云母花岗岩以及研究区全区进行39种元素R型聚类分析,得到地层、岩体以及全区的元素组合,但是分组的结果表明,通过R型聚类分析得到的各元素组合均不能明确反映地层、岩体以及全区的地质特征,因此对于R型聚类分析的使用需谨慎。     

区域化探数据处理的几种分形方法

地球化学变量兼有随机性、确定性和区域结构性的特征,传统的数学方法难以精确地描述地球化学变量的空间分布规律,而新兴的分形几何是刻画空间不规则形体的一种比较有效的工具,且分维值可以表征空间不规则形体的确定性本质。基于分形原理,提出含量-总量法、空间分形插值法和分形趋势面法3种新的分形技术,应用于区域化探数据处理。在几个典型矿区应用后发现,上述分形方法不仅可以有效地模拟地球化学变量的空间分布特征,而且在保留原始有效异常信息的基础上可以更合理地区分地球化学变量的背景与异常,具有发现和强化弱异常信息的重要功效。     

生态地球化学:从调查实践到应用理论的系统工程

生态地球化学是以多目标区域地球化学调查为基础,以生态地球化学评价、生态地球化学评估、生态地球化学预警和生态地球化学修复为主体的系统工程。文章系统地介绍了我国近年来生态地球化学研究成果,对长江流域河流生态系统、山西农田生态系统,北京城市生态系统,鄱阳湖、洞庭湖、太湖等湖泊湿地生态系统的区域生态地球化学评价进展进行了归纳总结。文章从区域尺度上综述了土壤碳库,土壤元素基准值和土壤背景值,土壤污染等级划分等基础研究方面的进展。面向未来,指出了生态地球化学发展趋势,提出生态地球化学研究需要进一步延伸和发展的领域。     

Statistical regression applied to borehole strain measurements data analysis

Summary In mining and geotechnical engineering, it is usually necessary to carry out field measurements in order to obtain information. Parameters are often measured indirectly and calculated based on certain relationships to the measured quantities. More often, the number of measurements taken is greater than the minimum required, in order to increase the reliability of results. However, some data points are less reliable than others for reasons such as measurement errors; a solution which best fits the measurement data is obtained accordingly. As a result, there is a residual or a difference between the individual quantities measured and those predicted from the best-fit solution. This brings about a question of how big a residual is acceptable for a solution to be reliable. It is also important to know whether the data point with the largest residual is the most erroneous, whether those data points with large residuals should be deleted and how many of them should be deleted. Standard deviation may provide a measure of the data divergence but it is questionable if this parameter can be used as a measure of the reliability of solution. In order to solve these problems, the author has done extensive study in this area, especially as part of geotechnical data analysis. In this paper, the statistical multiple regression method is introduced to analyse the measurement data. The method is applied to the analysis ofin situ stress measurement and can be easily adopted to analyse data from other field measurements and laboratory tests. An example is included which illustrates the analysis procedure and shows the advantages of the method.     

MATPLOT:一款基于MATLAB的地球化学数据分析成图软件

地球化学数据统计分析图解是目前岩石地球化学研究工作中常用的手段。目前常用的软件或多或少都存在一些不足之处。本文提出的MATPLOT是一款应用MATLAB编写而成的地球化学数据成图软件,可以在Windows/Mac OS/Linux多种操作系统平台中运行。与现有常用软件相比,该软件更具优势：可以更方便地实现较大地球化学数据量的导入,无需对数据进行繁琐的格式整理;更便捷地对数据进行二元散点图、三角图、标准化折线图、柱状图、概率密度图和同位素演化图等6种常用图解类型的成图,以及对数据图的展示内容和效果进行各种调整。此外,MATPLOT还可以对数据进行多种平均值计算（普通平均值、图基双权加权平均值和误差相关加权平均值）和相关性分析。

     

A comparison of regional geochemical data from lakes and streams in northern Labrador; implications for mixed-media geochemical mapping

Labrador has been covered by reconnaissance-scale geochemical surveys under the National Geochemical Reconnaissance program. Lake sediment and water were the chief sample media, but stream sediment and water were employed in the mountainous terrain of northern Labrador. The main objective of these surveys was mineral resource assessment, but the data are also relevant to geological and environmental studies, and would be most useful to the non-specialist if the data from the two drainage types could be combined to produce unified element distribution maps for the whole region.A comparison of stream and lake data for a 5,700-km² area where both drainage types were sampled suggests that only the pH of the lake and stream waters are directly comparable, showing a common range and similar spatial distribution. Comparing the two types of sediment, most elements show obvious differences in either median content or range or both, indicating that stream and lake sediment are geochemically distinct media, and their element contents cannot be compared directly. The distribution patterns of Cu, Ni and U reflect similar bedrock features in both sediment types. In contrast, Co, F, Fe, Hg, Mn, V and Zn show little or no spatial correlation between stream and lake sediment, but are strongly intercorrelated in the stream sediment data set.The sediments collected from lakes in Labrador represent disturbed column, about 40 cm in length, of organic debris that accumulated over the past several hundred years. Metal accumulation in the sediment is largely through fixation from inflowing surface and groundwater by microorganisms, coprecipitation with hydrous Fe and Mn oxides, sorption by clay minerals and chemical and biochemical processes at and just below the sediment/water interface. The stream sediments in this study were collected from active sediment, and represent principally the mechanical-weathering products of bedrock, with variable amounts of organic matter and hydrous Fe and Mn oxides. Considering the difference in the two sediment types, it is probably not surprising that there is rather limited spatial correlation between the geochemistry of the two sediment types indicating that to a large extent each medium reflects a different facet of the bedrock geochemistry. Only for a few elements should the data sets be merged. The degree of spatial correlation for U, Ni and Cu increases as the data are generalized by averaging into larger blocks, suggesting that the combined data sets will be more successful in defining broad crustal geochemical features rather than local details. The implication of this study for international geochemical mapping is that geochemical patterns for many elements are strongly dependent on the sample medium chosen. Therefore, when it is necessary to change sample media in passing from one terrain type to another, a comparative study must be carried out to determine how the geochemistry of the different sample media compare spatially.     

Kohonen neural network and factor analysis based approach to geochemical data pattern recognition

Kohonen neural network (KNN) and factor analysis are applied to regional geochemical pattern recognition for a Pb–Zn–Mo–Ag mining area around Sheduolong in Qinghai Province, China. Prior to factor analysis, the geochemical data are classified by KNN. The results demonstrate that the 4-factor model accounted for 67% of the variation in the data. Factor F1, a Pb–Zn–Mo factor and Factor F4, an Au–Ag factor, correlates with monzonitic granite intrusions and particularly with Pb–Zn–Mo–Ag mineralization within those rocks. Factor F2, an As–Co factor, correlates with metamorphic rocks of paleoproterozoic Baishahe formation. Factor F3, a Bi–Cu factor, correlates with granodiorite intrusions. The factor score maps suggest a revised location of faults and their mineralization significance in coarse geological map. The approach not only effectively interprets the geological significance of the factors, but also reduces the area of exploration targets.     

Cluster analysis of a regional-scale soil geochemical dataset in northern California

A regional-scale soil geochemical study was conducted within a 22,000 km² area in northern California including the Sierra Nevada, Sacramento Valley, and northern Coast Range. Over 1300 soil samples were chemically analyzed for 42 elements. The distribution of distinct groups of elements demonstrates the interplay of geologic, hydrologic, geomorphologic and anthropogenic factors; however, it is difficult to fully appreciate the complexity of geochemical transport and weathering processes on a landscape-scale in an area of very complex geology with such a large dataset containing more than 40 variables. To examine the data from a perspective of multi-element groupings, cluster analyses were applied to the dataset. The analysis identified several groups of elements whose spatial patterns could be related to specific geologic sources.     

ILR变换后数据的因子分区标准化在东昆仑东段地球化学异常圈定中的应用

从区域地球化学勘查数据中提取找矿地球化学信息是目前化探工作的重要研究课题之一。地球化学背景与异常划分是提取找矿地球化学信息的关键,而地球化学数据是成分数据,具有闭合效应,使得数据在进行主成分分析和因子分析时需进行变换。笔者采用东昆仑东段水系沉积物中11种常量元素的含量数据,通过ILR数据变换后进行因子分析,根据因子分区结果,结合EDA法,对样品数据进行分区标准化,以标准化后的值来圈定异常。结果表明,与传统方法相比,利用ILR变换后的因子分区标准化方法所圈定的异常可以去除地质体中高背景值的影响,与矿床空间位置对应较好,且该方法受人为因素影响较小,实际应用效果显著。     

Trace-element patterns in lateritic soils applied to geochemical exploration

Geochemical soil surveys in areas underlain by Precambrian volcano-metasedimentary sequences and around rare-metal-bearing pegmatites of southwestern Nigeria demonstrate that residual geochemical soil patterns reflect a wide range of potential source rocks adequately. The secondary geochemical dispersion processes in these typically tropical weathering environments adjust the trace-element distribution during lateritic soil development to narrow fluctuation ranges in comparison to the differing Clarke values of various source rock units.The sample density in these soil surveys, averaging at one sample per square kilometer, favours geochemical inventories even at regional scale and shows great potentials to predict bedrock composition of tropical terrain where rock outcrops are rather scarce.     

Factorial kriging analysis applied to geological data from petroleum exploration

A regionalized variable, thickness of the reservoir layer, from a gas field is decomposed by factorial kriging analysis. Maps of the obtained components may be associated with depositional environments that are favorable for petroleum exploration.     

利用区域化探数据推断地质体空间分布

区域化探数据可以反映地层的空间分布,利用区域化探数据借助有效的数据挖掘方法,能够提取出其中包含的地质信息,对于覆盖区填图以及矿产勘查有重要意义,其中的关键问题是如何进行数据挖掘。随机森林算法是近年来热门的机器学习方法,本文应用随机森林算法结合非平衡数据集分类方法提出了一种新的化探数据挖掘方法,通过实例研究验证表明该方法准确率高,能够有效的提取出区域化探数据中的地质信息。