Search

Mineral prospectivity mapping (MPM) is designed to reduce the exploration search space by combining and analyzing geological prospecting big data. Such geological big data are too large and complex for humans to effectively handle and interpret. Artificial intelligence (AI) algorithms, which are powerful tools for mining nonlinear mineralization patterns in big data obtained from mineral exploration, have demonstrated excellent performance in MPM. However, AI-driven MPM faces several challenges, including difficult interpretability, poor generalizability, and physical inconsistencies. In this study, based on previous studies, we devised a novel workflow that aims to constructing more transparent and explainable artificial intelligence (XAI) models for MPM by embedding domain knowledge throughout the AI-driven MPM, from input data to model design and model output. This newly proposed approach provides strong geological and conceptual leads that guide the entire AI-driven MPM model training process, thereby improving model interpretability and performance. Overall, the development of XAI models for MPM is capable of embedding prior and expert knowledge throughout the modeling process, presenting a valuable and promising area for future research designed to improve MPM.

Secondary carbonate minerals may influence the porosity of rocks; and, therefore, the quality of hydrocarbon and geothermal reservoirs. These carbonate minerals can be identified using shortwave infrared (SWIR) spectroscopy by their typical absorption features nearly $2.35~\mu \text{m}$ . However, their detection in quartz–illite–muscovite-dominated reservoir rock is hampered by the overlap of diagnostic features of carbonate minerals with the absorption features of illite–muscovite minerals. In this study, we used longwave infrared (LWIR) spectra to overcome this problem. We investigated the 6.50- and 11.30- $\mu \text{m}$ carbonate features in the LWIR as well as the 2.50- $\mu \text{m}$ feature at the longer wavelength margin of the SWIR to determine whether these features provide a better discrimination of carbonates in illite–muscovite-bearing rocks. We also investigated the possibility to use these features to quantify the carbonate content. We tested this on laboratory SWIR and LWIR spectra of 22 drill core samples from a Dutch reservoir sandstone of Permian age that contain variable amounts of illite–muscovite and carbonate minerals. Spectral parameters, such as the wavelength position and depth of absorption features (or reflectance peaks), were compared with (semi-) quantitative analyses of the mineral composition acquired with quantitative evaluation of minerals by scanning electron microscopy (QEMSCAN). Our results show that the ratio of band depths of the carbonate features at 2.44 and $2.50~\mu \text{m}$ are correlated with the relative carbonate content [carbonates versus illite + muscovite ratio (CIM)] with Spearman correlation coefficient ( $\rho$ ) of −0.87. The best predictor for the presence and content of carbonate minerals is the 6.50- $\mu \text{m}$ feature, with a correlation between the peak height and the absolute total carbonate content of 0.93. Therefore, the 6.50- $\mu \text{m}$ peak height can be best used for the quantification of carbonates in the quartz–illite–muscovite-dominated rocks we have investigated. The results of this study are useful for porosity studies of quartz–illite–muscovite rich reservoir rock, where carbonate minerals impact the rock porosity.

The systematic study of extreme geological events (such as plate collision and subduction, extreme cold and extreme hot events, biological extinction and revival, earthquakes, volcanoes, mineralization, and oil accumulation) that occurred during the evolution of the earth is essential not only for understanding the “abrupt changes in the evolution of the earth”, but also for an in-depth understanding of the co-evolution of material-life-environment of the livable earth. However, due to the temporal and spatial anomalies and complexity of extreme geological events, classical mathematical models cannot be effectively applied to quantitively describe such events. After comparative studies of many types of geological events, the author found that such extreme geological events often depict “singular” characteristics (abnormal accumulation or depletion of matter or massive release or absorption of energy in a small space or time interval). On this basis, the author proposes a unified definition of extreme geological events, a new concept of “fractal density” and a “local singularity analysis” method for quantitative description and modeling of extreme geological events. Applications of these methods to several types of extreme geological events have demonstrated that the singularity theory and methods developed in the current research can be used as general approaches for the characterization, simulation, and prediction of geological events.

Since the beginning of the 21st century, the geoscience research has been entering a significant transitional period with the establishment of a new knowledge system as the core and with the drive of big data as the means. It is a revolutionary leap in the research of geoscience knowledge discovery from the traditional encyclopedic discipline knowledge system to the computer-understandable and operable knowledge graph. Based on adopting the graph pattern of general knowledge representation, the geoscience knowledge graph expands the unique spatiotemporal features to the Geoscience knowledge, and integrates geoscience knowledge elements such as map, text, number, to establish an all-domain geoscience knowledge representation model. A federated, crowd intelligence-based collaborative method of constructing the geoscience knowledge graph is developed here, which realizes the construction of high-quality professional knowledge graph in collaboration with global geoscientists. We also develop a method for constructing a dynamic knowledge graph of multi-modal geoscience data based on in-depth text analysis, which extracts geoscience knowledge from massive geoscience literature to construct the latest and most complete dynamic geoscience knowledge graph. A comprehensive and systematic geoscience knowledge graph can not only deepen the existing geoscience big data analysis, but also advance the construction of the high-precision geological time scale driven by big data, the compilation of intelligent maps driven by rules and data, and the geoscience knowledge evolution and reasoning analysis, and others. It will further expand the new directions of geoscience research driven by both data and knowledge, break new ground where geoscience, information science, and data science intersect, and realize the original innovation of the geoscience research and achieve major theoretical breakthroughs in the spatiotemporal big data research.

Data-driven techniques including machine-learning (ML) algorithms with big data are re-activating and re-empowering research in traditional disciplines for solving new problems. For geoscientists, however, what matters is what we do with the data rather than the amount of it. While recent monitoring data will help risk and resource assessment, the long-earth record is fundamental for understanding processes. Thus, how big data technologies can facilitate geoscience research is a fundamental question for most organizations and geoscientists. A quick answer is that big data technology may fundamentally change the direction of geoscience research. In view of the challenges faced by governments and professional organizations in contributing to the transformation of Earth science in the big data era, the International Union of Geological Sciences has established a new initiative: the IUGS-recognized Big Science Program. This paper elaborates on the main opportunities and benefits of utilizing data-driven approaches in geosciences and the challenges in facilitating data-driven earth science transformation. The main benefits may include transformation from human learning alone to integration of human learning and AI, including ML, as well as from known questions seeking answers to formulating as-yet unknown questions with unknown answers. The key challenges may be associated with intelligent acquisition of massive, heterogeneous data and automated comprehensive data discovery for complex Earth problem solving.

Accessing the cyclicity and persistence of geological records with time series analysis deepens our understanding of Earth's long‐term evolution. In this study, time series from recent global zircon U‐Pb age and δ18O databases are analyzed using wavelet transform from a fractal perspective. Continuous wavelet transform of the local fractal sequences of both zircon records indicates a strong, persistent ~760 million year cycle over the 4.4 billion years of Earth's history. Wavelet coherence analysis shows that the U‐Pb age and δ18O systems are coupled in significant in‐phase coherence for a cycle of ~760 million year, implying synchronization between the two underlying processes. This study also demonstrates that the variation in the time series records manifests as 1/fβscaling behavior that persists β~ 1.8 over the entire interval. This 1/ffractal scaling nature furnishes evidence for the conjecture that rather than being an equilibrium system, Earth's long‐term evolution follows a self‐organized pattern. Determining the cyclicity and persistence of Earth's long‐term evolution can provide important clues for understanding the formation and growth of the continental crust. Significant obstacles to address the above questions include a lack of adequate data for global geological records over the course of Earth history and the sophistication of the time series analysis of these data. The recently improved global database of ~670,000 U‐Pb ages and ~6,000 δ18O isotope analyses from zircon grains allows the periodicity analysis of geological time series by a popular method of wavelet transform, form a fractal perspective. We found that the U‐Pb age and δ18O isotope systems are coupled in significant in‐phase coherence with a strong persistent ~760 Ma cycle throughout geological time ~4.4 Ga. We also demonstrate the harmonically (1/f) fractal nature of zircon time series, which provides evidence for the conjecture that the Earth's long‐term evolution is a self‐organized, far from equilibrium system. A persistent ~760 Ma cycle is identified in global zircon records across 4.4 GaU‐Pb age and δ18O systems are coupled in in‐phase coherence for an ~760 Ma cycleEarth's long term evolution is demonstrated with 1/ffractal nature

Fractal or multifractal concepts have significantly enlightened our understanding of crustal heterogeneity. Much attention has focused on 1/fscaling natures of physicochemical heterogeneity of Earth crust from fractal increment perspective. In this study, fractal density model from fractal clustering point of view is used to characterize the scaling behaviors of heterogeneous sources recorded at German Continental Deep Drilling Program (KTB) main hole, and of special contribution is the local and global multifractal analysis revisited by using Haar wavelet transform (HWT). Fractal density modeling of mass accumulation generalizes the unit of rock density from integer (e.g., g/cm3) to real numbers (e.g., g/cma), so that crustal heterogeneities with respect to source accumulation are quantified by singularity strength of fractal density in a-dimensional space. From that perspective, we found that the bulk densities of metamorphic rocks exhibit fractal properties but have a weak multifractality, decreasing with the depth. The multiscaling natures of chemical logs also have been evidenced, and the observed distinct fractal laws for mineral contents are related to their different geochemical behaviors within complex lithological context. Accordingly, scaling distributions of mineral contents have been recognized as a main contributor to the multifractal natures of heterogeneous density for low-porosity crystalline rocks. This finally allows us to use de Wijs cascade process to explain the mechanism of fractal density. In practice, the proposed local singularity analysis based on HWT is suggested as an attractive high-pass filtering to amplify weak signatures of well logs as well as to delineate microlithological changes. Fractal density modeling of physicochemical features in KTB deep holeFractal characterization of crustal heterogeneity for identifying lithological changesRevisiting multifractal density analysis by wavelet transform

Abstract: Mylonite is the result of the dynamic metamorphism and minerals in mylonite are deformed gradually with an increase in the degree of metamorphism. Quantifying the degree of deformation including the irregularities of shapes and the frequency distribution of the minerals becomes one of the most challenging efforts in mylonite analysis. Fractal modelling has been demonstrated in this paper to be an effective mean to achieve the above goal. Perimeter-Area fractal model was used to quantify the irregularities in the geometries and Cumulative Number-Area model is used to characterize the irregularities of distribution of quartzs in mylonites, respectively. Examples of quartz from five types of mylonites with different degree of deformation within the foreland of the Moine Thrust Zone in NW Scotland are chosen to study the evolution processes of deformation. As the main mineral component of quartzite mylonite, patterns are extracted from digital photomicrographics of the multiscale-grey image grid data to show quartz grains with different degree of deformation, The areas and perimeters of the quartz grains were calculated by GIS-based image processing technologies. From type one to type five, with an increase in degree of deformation, the corresponding Perimeter-Area exponent $${D_{\rm AP}}$$ increases from 1.20, 1.28, 1.38, 1.46, to 1.60, respectively, the fractal dimension $${D_{\rm P}}$$ of the perimeter from 1.07, 1.08, 1.17, 1.23, to 1.44, as well as the exponent of Cumulative Number- Area from 0.50, 0.51, 0.58, 0.82, to 0.85, respectively. The result has shown that as increase of the intensity of deformation, the shape of quartz grains tends to be more irregular, grain size tends to be smaller, and the number of grains increases. The results obtained using GSI model has indicated that as an increase in the intensity of deformation, the patterns of quartz grains tends to be more stratified and randomness increases.

The use of traditional statistical and new methods of spatial analysis can provide suitable indicators of geochemical element dispersion, and aid in targeting potential areas for mineral exploration. A case study involving the analysis of stream sediment geochemistry data from an area in Yunnan province, South China, is presented. The area has two known areas of gold deposits where several mines are present. The study starts with an appraisal of the descriptive statistics of the data. Geological studies were conducted in the study area to obtain a thorough understanding of the regional geology, ore geology of known gold deposits, and mineralization, and to determine the mineral deposit model. Most of the GIS analysis was done using the stream sediment geochemical data. An inverse distance weighting interpolation algorithm was used to convert the point data to continuous surface (grid) maps for each element. Principal component analysis (PCA) was used to compress the information to a few maps and to assist in determining multi-element associations. The study revealed that most of the high element concentrations in stream sediments were found in the Ailaoshan metamorphic belt. Gold deposits were found to be associated with ultramafic intrusives within the Ailaoshan metamorphic belt and the ultramafic instrusives are associated with principal component images that represent multi-element associations related to gold mineralization. The first two principal components possibly represent two different types of gold associations and phases of mineralization. The study demonstrates the usefulness of applying PCA to geochemical data to produce maps that reveal different associations useful for gold exploration.

Combined geostatistical and multifractal power-spectrum modelling of geochemical distributions can provide suitable indicators of metal dispersion, and is capable of analysing complex problems for targeting potential areas for mineral exploration. A case study analysing lake sediment geochemical data for the Gowganda area is presented and development of the methodology for spatial analysis of the data is described. The Gowganda-Cobalt area of northeastern Ontario is a textbook example of Co, Ag-Co vein-type deposit, which by 1984 had yielded one-half billion ounces of Ag. The area is also known for shear-zone-hosted Au mineralization. This paper uses the spatial and geometric distribution of lake sediment data to discriminate geochemical anomalies from background values. The application of two geostatistical techniques (spatial principal component analysis and indicator kriging) allows the estimation of geochemical distributions by utilizing their statistical and spatial properties. The newly developed multifractal power-spectrum method additionally allows for the geochemical distributions to be modelled by their multifractal Fourier-transformed power-spectrum characteristics. Verification of the estimates produced by these techniques has been enabled through spatial analysis of bedrock geology and mineral deposit occurrences in the area.

Current barriers hindering data-driven discoveries in deep-time Earth (DE) include: substantial volumes of DE data are not digitized; many DE databases do not adhere to FAIR (findable, accessible, interoperable and reusable) principles; we lack a systematic knowledge graph for DE; existing DE databases are geographically heterogeneous; a significant fraction of DE data is not in open-access formats; tailored tools are needed. These challenges motivate the Deep-Time Digital Earth (DDE) program initiated by the International Union of Geological Sciences and developed in cooperation with national geological surveys, professional associations, academic institutions and scientists around the world. DDE’s mission is to build on previous research to develop a systematic DE knowledge graph, a FAIR data infrastructure that links existing databases and makes dark data visible, and tailored tools for DE data, which are universally accessible. DDE aims to harmonize DE data, share global geoscience knowledge and facilitate data-driven discovery in the understanding of Earth's evolution.Deep-time Digital Earth is to promote data-driven discovery by harmonizing deep-time Earth data and creating ‘Geological Google’.

In general, the multifractal model provides more information about measurements on spatial objects than a fractal model. It also results in mathematical equations for the covariance function and semivariogram in spatial statistics which are determined primarily by the second-order mass exponent. However, these equations can be approximated by power-law relations which are comparable directly to equations based on fractal modeling. The multifractal approach is used to describe the underlying spatial structure of De Wijs 's example of zinc values from a sphalerite-bearing quartz vein near Pulacayo, Bolivia. It is shown that these data are multifractal instead of fractal, and that the second-order mass exponent (=0.979±0.011 for the example) can be used in spatial statistical analysis.

Perimeters and areas of similarly shaped fractal geometries in two-dimensional space are related to one another by power-law relationships. The exponents obtained from these power laws are associated with, but do not necessarily provide, unbiased estimates of the fractal dimensions of the perimeters and areas. The exponent (DAL) obtained from perimeter-area analysis can be used only as a reliable estimate of the dimension of the perimeter (DL) if the dimension of the measured area is DA=2. If DA<2, then the exponent DAL=2DL/DA>DL. Similar relations hold true for area and volumes of three-dimensional fractal geometries. The newly derived results are used for characterizing Au associated alteration zones in porphyry systems in the Mitchell-Sulphurets mineral district, northwestern British Columbia.

A fuzzy relation analysis method is used to derive weights for qualitative variables based on their partial order relations. Two asymmetric measure indexes (incidence coefficient and probability difference) are proposed to measure the asymmetric associations between geoscience variables from which the partial order relations can be constructed. The fuzzy relation analysis method can be implemented in combination with the asymmetric measure indexes leading to new methods for pattern overlay and data integration in mineral potential prediction. Two types of models are proposed and illustrated by two artificial examples: one for predicting targets for undiscovered deposits, and the other for estimating the mineral resource potential of the targets.

Concepts of fractal/multifractal dimensions and fractal measure were used to derive the prior and posterior probabilities that a small unit cell on a geological map contains one or more mineral deposits. This has led to a new version of the weights of evidence technique which is proposed for integrating spatial datasets that exhibit nonfractal and fractal patterns to predict mineral potential. The method is demonstrated with a case study of gold mineral potential estimation in the Iskut River area, northwestern British Columbia. Several geological, geophysical, and geochemical patterns (Paleozoic-Mesozoic sedimentary and volcanic clastic rocks; buffer zones around the contacts between sedimentary rocks and Mesozoic intrusive rocks; a linear magnetic anomaly; and geochemical anomalies for Au and associated elements in stream sediments) were integrated with the gold mineral occurrences which have fractal and multifractal properties with a box-counting dimension of 1.335±0.077 and cluster dimension of 1.219±0.037.

The interrelationships between two previously developed multifractal models are discussed. These are the Evertsz-Mandelbrot model developed on the basis of the multifractal spectrum f(α), and the Schertzer-Lovejoy model based on the codimension function C(γ) where α and γ represent Hölder exponent and field order, respectively. It is shown how these two models are interrelated: they are identical for values of γ within the range D−α(0)≤γ≤D−αmin. where D is the Euclidean dimension. For D−αmax≤γ≤D−α(0), however, f(α) remains a continuous function of α whereas C(γ) assumes constant value. In this respect, the fractal spectrum f(α) can provide more information about the multifractal measure than the codimension function C(γ). The properties of the two models are illustrated by application to the binomial multiplicative cascade model.