Identifying Weak But Complex Geophysical And Geochemical Anomalies Caused By Buried Ore Bodies Using Fractal And Wavelet Methods

The past decades have seen that targeting and prospecting the concealed ore deposits especially for the huge/super-huge deposits in covered area (e.g., desert, regolith and vegetation) becomes one of the most attractive but challenged carrers for mineral exploration and resource assessment. Mineral resource quantitative prediction and assessment not only need to investigate the ore genesis, determine the correct indicators for prospecting ore deposit and building the rational model for mineral resource prediction, but also necessitate the information extraction and assessment methods to identify geo-anomalies of interests (e.g., geophysical and geochemical anomalies, remote sensing information, etc) related to mineralization. However, extracting and identifying weak but complex geo-anomalies caused by the deeply buried ore bodies bring the crucial challenges for finding the potential mineral resources in covered area. The real geo-anomaly patterns generally exhibit very irregular signatures with both structured and random components, and recent studies evidence that these patterns manifest both multiscale and scale-invariant properties as a natural consequent of hierarchies of earth and mineralization system. Nonlinear science and complexity theory provide the contexts of fractals/multifractals (e.g., fractal dimension, scale invariance, selfsimilarity, singularity, etc.) to describe, model and analyze the deep regularitied buried in complex pattern. Numerous fractal models have been proposed to model and analysis the geochemical and geophysical fields, such as scaling (1/f) noise model, concentration-area (C-A) model, spectrum-area (S-A) model and fractal density model, to name but few examples. These kinds of fractal analysis technique have been significantly facilitated the identification of geoanomalies assisting in mineral exploration. In fact, fractal analysis such as the box counting method is naturally related to the concept of multi-resolution which needs to investigate the multiscale natures of objects. The wavelet transform, as the most famous, powerful and advanced multiscale analysis tool, have been early suggested as a natural tool for fractal/multifractal analysis, but not yet been paid much attentions to multiscale analyzing the geoanomalies especially for scaling/multiscaling analysis for singular geochemical distribution and geophysical field. In this thesis, after reviewing the theories of fractal/multifractal and and methodologies of wavelet-based multiscale analysis, 1/f scaling nosie model is firstly introduced to investigate the scaling natures as well as the mechanism of petrophysical distributions (e.g., metals, susceptibility, density) within the earth crust. Meanwhile, special intentions have been paied to improve the conventional model of data processing and interpretation in geophysical and geochemical exploration by using 1/f scaling model, including the proposed fractal-based matched filtering method for magnetic anomalies separation, the missed data predcition and the improved S-A filtering method for geochemical data analysis. In particular, the subsequent chapters of this thesis are devoted to elaborate on the ideas and methods of singularity analysis and multifractal filtering developed in the context of multifractal based on wavelet transform for analyzing and identifying weak but complex anomalies caused by buried mineralization. Firstly, the generalized fractal density model is presented using wavelets, and we argue that the proposed wavelet-based singularity analysis improves the identification of the weak geochemical/geophysical anomalies compared with conventional moving-averaging-based LSA, and the subsequent sessions are followed by introduction of a novel method for multifractal spectrum calculation using wavelet-based fractal density model. Subsequently, the statisitic distribution of wavelet coefficients has been investigated by using fractal/multifractal model. In light of the fractal/multifractal model outperforming the conventional (generalized) Gaussian model for characterizing wavelet coefficient statistical distribution, a novel multifractal filtration has been outlined for separating the mixed geo-information, such as the separation of background and anomaly, regional and residual components. Throughout this thesis, two case studies from metallogenic districts including Inner Mongonia in North China and Nanling Range in South China are used to validate the proposed theories and methods, including wavelet-based matched filtering, singularity analysis and multifractal filtering, for identifying geophysical/geochemical anomalies caused by buried mineral deposits or underlying subsurface geologies. As a result, the facts demonstrate that these proposed nonlinear methods and techniques based on fractal and wavelet theories have potential to be powerful tool assisting in mineral exploration, especifically for dealing with the two major challenges encountered in the covered area including enhancing the weak but complicated geoanomalies and separating the mixed geoinformation.