Research on processing geochemical data and identifying geochemical anomalies has made important progress in recent decades. Fractal/multi-fractal models, compositional data analysis, and machine ...learning (ML) are three widely used techniques in the field of geochemical data processing. In recent years, ML has been applied to model the complex and unknown multivariate geochemical distribution and extract meaningful elemental associations related to mineralization or environmental pollution. It is expected that ML will have a more significant role in geochemical mapping with the development of big data science and artificial intelligence in the near future. In this study, state-of-the-art applications of ML in identifying geochemical anomalies were reviewed, and the advantages and disadvantages of ML for geochemical prospecting were investigated. More applications are needed to demonstrate the advantage of ML in solving complex problems in the geosciences.
The Gangdese Belt is now recognized as an important Cu polymetallic mineralization belt. Recent studies suggest that apart from porphyry copper deposits, polymetallic skarn deposits are another ...significant deposit type in this belt. In this paper, principal component analysis (PCA) and spectrum–area (
S–
A) fractal modeling are used to identify geochemical anomalies associated with Cu and Pb–Zn skarn mineralization based on Cu, Pb, Zn and Ag stream sediment data. Firstly, the raster maps of Cu, Pb, Zn and Ag were obtained by multifractal inverse distance weighted (MIDW). Secondly, PCA was used to combine the Cu, Pb, Zn and Ag concentration values. Finally,
S–
A analysis was used to decompose the first component pattern obtained by the PCA. The results show that major anomalies of Cu, Pb, Zn and Ag are mostly located around mapped intrusions, and along E–W trending faults, where Cu and Pb–Zn skarn deposits favorably occur. These results indicate that: (1) the Gangdese Belt potentially hosts undiscovered Cu and Pb–Zn skarn deposits, and the integrated anomalies for Cu, Pb, Zn and Ag occurring around the intrusions and in the vicinity of faults in the center part of the Gangdese Belt should be further investigated in the next step of mineral resource exploration; and (2) the hybrid method combining PCA and
S–
A modeling is an effective tool to identify geochemical anomalies.
► Geochemical mapping plays an important role in mineral resource exploration. ► Spectrum–area fractal model is an effective tool to decompose mixed geochemical pattern. ► Gangdese Belt potentially hosts undiscovered skarn deposits.
► The spatial distribution of arsenic in Gangdese belt shows a complex and mixed pattern. ► The spectrum-area fractal model (S-A) is a powerful tool to decompose a mixed pattern. ► The arsenic ...anomalies may related to Cu mineralization because the areas with high anomaly of arsenic mainly are located surrounding known Cu deposits.
Decomposing mixed geochemical patterns is a challenge in geochemical exploration and environmental assessment. In this paper, the spectrum–area technique (
S–
A) is used to decompose a mixed pattern of arsenic in Gangdese belt based on stream sediment data.
S–
A is a multifractal model based on power–law relationships between area of the set consisting of wave numbers with spectral energy density above
S
A(>
S) on the 2D frequency domain. The original spatial distribution map of arsenic obtained by inverse distance weighted (IDW) shows a mixed pattern due to superposition of different geological processes or events and is converted into the frequency domain by means of Fourier transformation. Two components, including power spectrum density and phases, are obtained. The spectrum energy density (
S) and the area (
A) enclosed by the above-threshold spectrum energy density is plotted on a log–log scale. Two cutoff values determined by three straight lines define three filters which decompose the original map of arsenic into background, anomalous, and high frequency (noise) components. The areas with high anomaly of arsenic mainly are located surrounding known Cu deposits, indicating that arsenic anomalies may be related to Cu mineralization.
Machine learning algorithms have been applied widely in the fields of natural science, social science and engineering. It can be expected that machine learning approaches especially deep learning ...algorithms will help geoscientists to discover mineral deposits through processing of various geoscience datasets. This study reviews the state-of-the-art application of deep learning algorithms for processing geochemical exploration data and mining the geochemical patterns. Deep learning algorithms can deal with complex and nonlinear problems and, therefore, can enhance the identification of geochemical anomalies and the recognition of hidden patterns. Applied geochemistry needs more applications of machine learning and/or deep learning algorithms.
•The application of deep learning algorithms for geochemical mapping was reviewed.•The advantages of deep learning algorithms in geochemical mapping was demonstrated.•Challenges and future work on deep learning algorithms in geochemical mapping were discussed.
The Fanshan district is a prospective area to explore for epithermal-type Cu–Au mineralization that is similar to the Zijinshan Cu–Au deposit in the neighboring region which is the largest Cu–Au ...deposit in southern China. In this study, the robust principal component analysis (RPCA) method and spectrum-area (S-A) fractal model were applied to multi-element data from 2042 stream sediment samples collected in this region. The isometric logratio (ilr) transformation was used to open the geochemical data for reducing the effects of data closure problem. The resulting PC1 obtained using RPCA displays two different compositional assemblages: (I) Pb, Zn, Sn, W, Mo, Bi, Hg, and Ag, perhaps representing hydrothermal-type Zn–Pb polymetallic mineralization, and (II) As, Au, Cu, Sb and Mn, likely representing epithermal-type Cu–Au mineralization. The decomposed anomaly map from the PC1 obtained from the S-A model illustrates that highly anomalous areas occurring around the intrusions are correlated with hydrothermal-type Zn–Pb polymetallic mineralization, and areas with lower anomalies such as Dafanshan and Xiaofanshan are perhaps associated with Cu–Au polymetallic mineralization, and should be further studied to explore for epithermal-type Cu–Au mineralization.
•Fanshan district is a prospective area to explore for Cu–Au mineralization.•The data closure problem is addressed.•The robust principal component analysis was used to combine multiple variables.•The resulting of RPCA shows two different composition assemblages.•The mixed geochemical pattern is decomposed by the S-A model.
Over the past several decades, a wide range of complex structures or phenomena of interest to geologists and geochemists has been quantitatively characterized using fractal/multifractal theory and ...models. With respect to the application of fractal/multifractal models to geochemical data, the focus has been on how to decompose geochemical populations or quantify the spatial distribution of geochemical data. A variety of fractal/multifractal models for this purpose have been proposed on the basis of the scaling characteristics of geochemical data. These include the concentration–area (C-A) fractal model, concentration–distance (C-D) fractal model, spectrum–area (S-A) multifractal model, multifractal singularity analysis, and the concentration–volume (C-V) fractal model. These fractal models have been widely demonstrated to be useful, as indicated by the increasing number of published papers. In this study, fractal/multifractal modeling of geochemical data including its theory, the way it works, its benefits and limitations, its applications, and the relationships between these models are reviewed. The comparison among of C-A, S-A, and multifractal singularity analysis based on simulated data suggested that mapping singularity technique can enhance and identify weak anomalies caused by buried sources. Future study should focus on how to distinguish the true anomalies associated to mineralization with the false anomalies from a fractal/multifractal perspective.
•Fractal/multifractal modelling of geochemical data is reviewed.•The C-A, S-A and singularity index are compared based on simulated data.•Singularity mapping technique can well detect weak geochemical anomalies.
In this paper, we train an autoencoder network to encode and reconstruct a geochemical sample population with unknown complex multivariate probability distributions. During the training, small ...probability samples contribute little to the autoencoder network. These samples can be recognized by the trained model as anomalous samples due to their comparatively higher reconstructed errors. The southwestern Fujian district (China) is chosen as a case study area. A variety of learning rates, iterations, and the size of each hidden layer are constructing and training the deep autoencoder networks on all the geochemical samples. The reconstruction error (or, anomaly score) of each training sample is used to recognize multivariate geochemical anomalies associated with Fe polymetallic mineralization. By comparing the results obtained with a continuous restricted Boltzmann machine, we conclude that the autoencoder network can be trained to recognize multivariate geochemical anomalies. Most of the known skarn-type Fe deposits are located in areas with high reconstruction errors or anomaly scores in the anomaly map, indicating that these anomalies may be related to Fe mineralization.
•Recognition of geochemical anomalies using an autoencoder network.•Recognition of geochemical anomalies using continuous restricted Boltzmann machines.•Methods demonstrated in a case study southwestern Fujian district (China).
The recognition of multivariate geochemical anomalies is important for mineral exploration. Big data analytics, which involves the whole data and variables, is an alternative manner to delineate ...multivariate geochemical anomalies in support of machine learning algorithms due to their strong ability to capture the complex intrinsic and diverse links between geochemical characteristics and mineralization. However, this method faces the issue of data redundancy and calculation complexity, and high-dimensional problems raise great challenges for anomaly detection. This is the curse of dimensionality problem, which hinders the development of a variety of techniques for anomaly detection. In this study, a hybrid model that combines unsupervised deep belief networks (DBNs) and one-class support vector machine (OCSVM) is adopted to address the high-dimensional geochemical anomalies detection problem. In this model, the relevant features first extracted through the DBN are used as the input of the OCSVM. The decision function values of the hybrid method are employed to map the geochemical patterns related to iron mineralization. The comparative results on the performance of the hybrid model and the other three anomaly detection models (deep autoencoder model, OCSVM, and hybrid model with principal component analysis and OCSVM) in terms of the area under curve(AUC) values, suggest that the hybrid method of the DBN and OCSVM can efficiently recognize the geochemical anomalies related to iron mineralization. The DBN can extract the geochemical information, reduce the redundant features, and further enhance the scalability of the OCSVM for processing high-dimensional geochemical data. The extracted geochemical anomalies, which show a close spatial relationship with the Yanshanian intrusions, can provide significant guidance for the next round of mineral exploration.
•A combination of a one-class SVM and deep learning for multivariate geochemical anomalies recognition is demonstrated.•The hybrid model delivers a better accuracy than the single one-class SVM for anomaly detection.•Deep learning can extract high-level features compared to linear feature extraction method.•A case study from southwestern Fujian district was conducted.
Mineral prospectivity mapping (MPM) aims to reduce the areas for searching of mineral deposits. Various statistical models that have been successfully adopted to delineate prospecting regions for a ...specific type of mineral deposit can be divided into pixel-wise, image- (or pixel-patch), and graph-based approaches. The pixel-wise models, which frequently integrate multiple prospecting information (or evidence layers) at a single pixel, do not adequately consider the spatial associations among neighboring pixels and may ignore the spatial patterns linked to mineralization or the spatial distribution characteristics of mineral deposits to some extent. Image-based models such as convolutional neural networks (CNNs) can extract local meaningful features and capture the spatial patterns of prospecting information in MPM because the input data of image-based models are images composed of regular pixels in the Euclidean space. However, CNNs also have limitations in MPM, such as the requirement for regular input data and non-rotationally invariant spatial features. Graphs that are typically composed of nodes and edges have a strong abstraction to capture the complex and nonlinear spatial relationships among multiple nodes and their edges. Prospecting information or evidence layers can be regarded as graphs in which pixels are connected by their adjacent pixels. In this study, graph deep learning algorithms, including graph convolutional networks and graph attention networks, were employed to produce mineral potential maps. A comparative study of graph deep learning algorithms with a CNN demonstrated the advantage of graph deep learning algorithms for MPM in terms of the cumulative areas versus the cumulative number of mineral deposits and the true/false prediction rate plot. These results suggest that the graph-based models, such as graph neutral networks, can effectively capture mineralization information and the spatial interrelations between mineralization and prospecting information.
Deep neural networks perform very well in learning high-level representations in support of multivariate geochemical anomaly recognition. Geochemical exploration data typically contain a proportion ...of large variations and missing values, which motivated us to construct a network architecture optimized to deal with these data. Our approach adopted a stacked convolutional denoising autoencoder (SCDAE) to extract robust features and decreased the level of sensitivity to partially corrupted data, that is, input data that are partially missing. SCDAE parameters, which include the network depth, number of convolution layers, number of convolution kernels, and convolution kernel size, were optimized using trial-and-error experiments. The optimal SCDAE architecture was then used to recognize multivariate geochemical anomalies related to mineralization in a case study in southwestern Fujian Province, based on the differences in the reconstruction errors between sample populations. The spatial distribution of high reconstruction errors in the anomaly map was closely related to most known Fe deposits, indicating the effectiveness of the SCDAE at recognizing geochemical anomalies related to Fe mineralization. A comparative study between the SCDAE and a stacked convolutional autoencoder (SCAE) with different corruption levels showed that the SCDAE exhibited reduced sensitivity to stochastic disturbances with different corruption proportions, and had an enhanced ability to recognize geochemical anomalies varying in a reasonable range. The robustness of the SCDAE makes it applicable to a wide variety of geochemical exploration scenarios, particularly in areas with incomplete or missing data.