Search

Effective management of surface emissions from abandoned gob methane (AGM) is crucial for mitigating greenhouse gas emissions and ensuring public safety. An important geochemical characteristic of near-surface AGM migration is the potential presence of abnormal component concentrations in near-surface soil gas over abandoned coal gobs. To investigate this phenomenon, a surface geochemical survey was conducted based on four survey lines in the Dongyi and Dongsan abandoned gob groups in the Xinghua Coal Mine, Jixi Basin, in China. The gas chromatography technique was used to analyze the concentrations of methane, carbon dioxide, ethane, propane, and butane in the collected 43 soil gas samples. The results revealed a significant anomaly of soil gas concentrations, particularly methane and carbon dioxide anomalies, in the near-surface soil over abandoned gobs. The background concentration for methane was determined to be 7.49 ppm, with an anomalous threshold set at 10 ppm based on a statistical analysis and an iterative method. This threshold could be confirmed by examining the coupling and decoupling relationship between methane, carbon dioxide, and C2–3 as well. A spatial correlation between regions exhibiting anomalous methane and carbon dioxide concentrations and the positions of gob areas, abandoned surface wells, and faults was observed. Abandoned and sealed coalbed methane surface wells and faults near gas-rich gob areas have the potential to act as conduits for AGM leakage to the surface. Furthermore, concentrations of methane, carbon dioxide, and C2–3 in soil gas over abandoned coal gobs were significantly higher compared to areas unaffected by mining activities. This suggests that elevated concentrations of methane, carbon dioxide, and C2–3 in soil gas may originate from underground AGM. [ABSTRACT FROM AUTHOR]

Regional geochemical surveys generate large amounts of data that can be used for a number of purposes such as to guide mineral exploration. Modern surveys are typically designed to permit quantification of data uncertainty through data quality metrics by using quality assurance and quality control (QA/QC) methods. However, these metrics, such as data accuracy and precision, are obtained through the data generation phase. Consequently, it is unclear how residual uncertainty in geochemical data can be minimized (denoised). This is a limitation to propagating uncertainty through downstream activities, particularly through complex models, which can result from the usage of artificial intelligence-based methods. This study aims to develop a deep learning-based method to examine and quantify uncertainty contained in geochemical survey data. Specifically, we demonstrate that: (1) autoencoders can reduce or modulate geochemical data uncertainty; (2) a reduction in uncertainty is observable in the spatial domain as a decrease of the nugget; and (3) a clear data reconstruction regime of the autoencoder can be identified that is strongly associated with data denoising, as opposed to the removal of useful events in data, such as meaningful geochemical anomalies. Our method to post-hoc denoising of geochemical data using deep learning is simple, clear and consistent, with the amount of denoising guided by highly interpretable metrics and existing frameworks of scientific data quality. Consequently, variably denoised data, as well as the original data, could be fed into a single downstream workflow (e.g., mapping, general data analysis or mineral prospectivity mapping), and the differences in the outcome can be subsequently quantified to propagate data uncertainty. [ABSTRACT FROM AUTHOR]

Geochemical surveys contain an implicit data lifecycle or pipeline that consists of data generation (e.g., sampling and analysis), data management (e.g., quality assurance and control, curation, provisioning and stewardship) and data usage (e.g., mapping, modeling and hypothesis testing). The current integration of predictive analytics (e.g., artificial intelligence, machine learning, data modeling) into the geochemical survey data pipeline occurs almost entirely within the data usage stage. In this study, we predict elemental concentrations at the data generation stage and explore how predictive analytics can be integrated more thoroughly across the data lifecycle. Inferential data generation is used to modernize lake sediment geochemical data from northern Manitoba (Canada), with results and interpretations focused on elements that are included in the Canadian Critical Minerals list. The results are mapped, interpreted and used for downstream analysis through geochemical anomaly detection to locate further exploration targets. Our integration is novel because predictive modeling is integrated into the data generation and usage stages to increase the efficacy of geochemical surveys. The results further demonstrate how legacy geochemical data are a significant data asset that can be predictively modernized and used to support time-sensitive mineral exploration of critical minerals that were unanalyzed in original survey designs. In addition, this type of integration immediately creates the possibility of a new exploration framework, which we call predictive geochemical exploration. In effect, it eschews sequential, grid-based and fixed resolution sampling toward data-driven, multi-scale and more agile approaches. A key outcome is a natural categorization scheme of uncertainty associated with further survey or exploration targets, whether they are covered by existing training data in a spatial or multivariate sense or solely within the coverage of inferred secondary data. The uncertainty categorization creates an effective implementation pathway for future multi-scale exploration by focusing data generation activities to de-risk survey practices. [ABSTRACT FROM AUTHOR]

Deep learning algorithms (DLAs) are becoming hot tools in processing geochemical survey data for mineral exploration. However, it is difficult to understand their working mechanisms and decision-making behaviors, which may lead to unreliable results. The construction of a reliable and interpretable DLA has become a focus in data-driven geoscience discovery. This study utilized a SHapley Additive exPlanations (SHAP) framework, a popular post-hoc interpretability analysis method, incorporated with the variational autoencoder (VAE) to explore the contribution of geochemical elements for multivariate geochemical anomaly recognition. The sorting of element importance obtained by SHAP tool can provide a novel view for selecting a suitable elemental association related to mineralization. Based on the metallogenic model in the southeastern Hubei Province of China, a metallogenic-factor-based VAE model was constructed using an ad-hoc interpretable modeling technique. The interpretability of the model in identifying the abnormal distribution of the element associations can be improved by constructing a hidden layer and loss function containing metallogenic regularity and key metallogenic factors. The highly anomalous areas identified by the metallogenic-factor VAE model not only contain most of the known Au deposits, but also can reasonably identify the abnormal elemental associations related to ore-forming processes under the guidance of the metallogenic regularity. According to the output visualization of the new hidden layer, and the results of receiver operating characteristic curve and success-rate curve, the metallogenic-factor VAE model exhibits satisfied interpretability and performance. The geochemical anomalies identified in this study provide critical clues for future mineral exploration. [ABSTRACT FROM AUTHOR]

Mapping of lithological units is a significant challenge for geological tasks. Stream sediment geochemical survey data contain abundant geological information that can help delineate lithological units. In this study, a convolutional neural network (CNN) was applied to map the lithological units in the Daqiao gold District, West Qinling Orogen, China, based on stream sediment geochemical data, in which each sample includes the concentrations of 15 trace elements (Cu, Pb, Zn, Ag, Mo, Sn, W, Mn, Ba, As, Sb, Bi, Cd, Au, and Hg). The training samples were firstly constructed with a certain window size by randomly selecting locations within each lithological unit. A CNN model was then established based on AlexNet to classify the lithologic categories. The classification map showed that 7 lithological units were correctly distinguished with an overall classification accuracy of 90.0%, suggesting that (1) stream sediment geochemical survey data of only trace element concentrations are useful for lithological mapping, and (2) a CNN can extract effectively geochemical characteristics from geochemical survey data. This study confirms the potential of a CNN as an effective method for geological mapping based on geochemical survey data. [ABSTRACT FROM AUTHOR]

Machine learning methods have recently been used widely for mineral prospectivity mapping. However, few studies have focused on the determination of variables for mineral prospectivity prediction using such methods. Here, we present a comparative study using supervised and unsupervised methods to determine predictive variables (PVs). First, based on a mineral deposit model, 12 variables were created including information about granite, fault and strata, and information from geochemical and geophysical surveys. Second, recursive feature elimination (RFE) and sparse principal components analysis (SPCA) were used to determine the PVs for mineral prospectivity prediction. Third, the weights-of-evidence and Random Forest methods were used to integrate the PVs to generate a probability map of mineral prospectivity. Finally, the receiver operating characteristic curve was used to evaluate the performance of the PVs for indicating mineral prospectivity. The variable strata buffer, granite buffer, stratigraphic entropy, derivative norm of magnetic data, and fault buffer were selected as PVs by SPCA, whereas the derivative norm of magnetic data, fault buffer, geochemical anomalies, and strata number were selected as PVs by the RFE method. The results demonstrate that PV determination is a necessary step for mineral prospectivity mapping because it can improve the performance of mineral prospectivity prediction. [ABSTRACT FROM AUTHOR]

To integrate the evidential signatures into a mineral prospective map, it is necessary to model the observed data of metallogenic predictors with complex relationships. In this paper, a methodology was developed for gold prospectivity modeling by combining Laplacian eigenmaps (LEMS) and the least angle regression (LARS). LEMS were used to transform metallogenic predictors into Laplacian spectra to enhance the data classification ability, and the Laplacian spectra were used in LARS for gold prospectivity modeling. The Jinchanggouliang area, an important gold metallogenic district in China, was taken as the study area. Based on the geological and geochemical survey data of the study area, the LARS model, the combination of LEMS and LARS (LEMS–LARS) model, the logistic regression (LGR) model and the one-class support vector machine (OCSVM) model were each established for gold prospectivity modeling in the study area. The performances of these four models were measured using different performance evaluation methods. The receiver operating characteristic (ROC) analysis reveals that the LEMS–LARS model, the LGR model and the OCSVM model yield comparable performances and were superior to the LARS model, for gold prospectivity modeling in the study area. The prediction–area (P–A) plots of the four models show that the LEMS–LARS model, the LGR model and the OCSVM model each predicted about 20% of the study area as gold prospective areas, which contain about 80% of the known gold deposit locations. However, the performance of the LARS model was difficult to evaluate by the P-A plot because its prediction curve overlaps with its area curve, forming a straight-line segment, in the P-A plot. The gold prospective areas delineated optimally separately by the LEMS–LARS model, the LGR model and the OCSVM model are associated spatially with areas having favorable metallogenic conditions. The case study illustrates that the LEMS–LARS model can be used as a data-driven mineral prospectivity method, and its performance, in gold prospectivity modeling in the study area, is comparable to that of the LGR method and OCSVM method. [ABSTRACT FROM AUTHOR]

Fractal and multifractal models, including the concentration-area (C–A) fractal model, spectrum-area (S–A) multifractal model, and local singularity analysis (LSA) method, are widely applied when processing various geoscience datasets. However, there is lack of ArcGIS-based software that contains these popular fractal and multifractal models. Such a situation hinders the popularization and application of fractal and multifractal models. ArcFractal, an easy-to-use ArcGIS add-in for processing geoscience data using fractal and multifractal models, is introduced in this paper. It is developed using C# based on ArcObject for.Net, ArcEngine 10.2, ZedGraph, and Visual Studio 2010. ArcFractal operations require a Windows 7 operating system and ArcGIS Desktop, version 10.2 or higher. The main application of ArcFractal is to determine geochemical threshold/baseline for separating geochemical patterns into anomalous and background components using the C–A, S–A, and LSA techniques. A case study from China Geochemical Baselines Project (CGB) was used to demonstrate the advantage of ArcFractal for processing geochemical exploration data. [ABSTRACT FROM AUTHOR]

The weighted mean and the multiple regression techniques are two methods that are employed to estimate elemental background concentration of lithologies upstream of stream sediment samples and the background value of elements at catchment basin outlets due to the lithology. However, these techniques have not previously been validated against real background values by sampling from lithological units. A geochemical survey was completed in central Iran by collecting samples from stream sediments along with rock samples from upstream lithologies. The accuracy and the precision of both techniques were examined graphically and statistically. The results demonstrated the advantage of multiple regressions over the weighted mean method. However, estimated background values from both techniques show high deviations from the analyzed values for some elements including Cr, Ni, Pb and Zn. While the weighted mean method is not as sensitive as the multiple regression approach to small areas of lithological units (independent variables), the estimated background values emerging from the weighted mean technique follow narrow ranges among different lithological units and cannot highlight the differences between elemental concentrations in various lithologies. However, multiple regression analysis is more successful than the weighted mean method in estimating expected background values of sediments at the catchment basin outlet. The multiple regression method uses the areas of all lithologies to estimate the background value of a particular lithology, whereas in the weighted mean method, only the area of the lithology is involved in the background estimation of that lithology. The existing equations essentially take into account only mechanical dispersion but overlook hydromorphic dispersion of elements in fluvial systems. Moreover, these equations only use area of rock units as the weighting factor or independent variables. Several parameters such as rate of weathering and erodibility, transportation mode, scavenging effects, the sediment delivery ratio of a catchment basin and each lithological unit control sediment chemistry at the basin outlet and should also be taken into account. [ABSTRACT FROM AUTHOR]

A variety of factors contribute to the proper and efficient evaluation of geochemical programs including sampling medium, sample size and sample fraction, sampling density, digestion procedure, and analytical techniques. Among the aforementioned factors, sampling density is of great importance, affecting project ability to detect anomalous sites, which are either economically viable or environmentally hazardous. It is quite critical to have a sufficient number of samples to delineate geochemical anomalies or baselines and to reveal the relative variations in elemental concentrations in a study area. The resolution of the subsequent maps obtained by sampling greatly depends on sampling density. While a large number of samples may generate redundant data, a small number of samples smooth out the patterns and may completely miss anomalies. This review provides a platform for applied geochemists who desire to design a sampling scheme for a regional geochemical program and do not have the financial resources to perform an orientation survey, which will give them the necessary information for designing, according to project objectives, a cost-effective regional geochemical survey for their study area. It is advisable to integrate during the desktop study all available information, such as geological, structural, geophysical, remote sensing data and topographical data by using different data fusion algorithms to arrive at a sampling density fit-for-purpose of the target area. This in turn facilitates an effective spatial positioning of sample locations. Quantitative analyses of the efficiency of different sampling densities for delineating geochemical patterns have not been well documented, warranting more attention and research. [ABSTRACT FROM AUTHOR]

Quantitative prediction of concealed mineralization is always confronted with difficulties in comprehensive analysis between 2D and 3D data and between qualitative and quantitative data. A weighted shortest-distance field method is proposed here to track, in 3D heterogeneous space, the shortest migration paths of ore-forming elements from an orebody to the ground surface, assuming that ore-forming elements migrate at less costs into fault rupture zones than in other surrounding rocks. This method was used to generate the weighted shortest-distance field of the 3D orebody model in the Hongtoushan copper mine, NE China. In addition, the field value and the geochemical soil survey data on the Earth's surface were subjected to statistical analysis. Results showed that some geochemical anomalies are characterized by the shortest-distance field of the known orebodies, while other formerly unrecognized anomalies may possibly be related to undiscovered orebodies. This method can also be applied to comprehensive statistical analysis between a 3D geological model and 2D data on the Earth's surface, e.g., geophysical exploration or remote sensing data. [ABSTRACT FROM AUTHOR]

Prediction of true classes of surficial and deep earth materials using multivariate spatial data is a common challenge for geoscience modelers. Most geological processes leave a footprint that can be explored by geochemical data analysis. These footprints are normally complex statistical and spatial patterns buried deep in the high-dimensional compositional space. This paper proposes a spatial predictive model for classification of surficial and deep earth materials derived from the geochemical composition of surface regolith. The model is based on a combination of geostatistical simulation and machine learning approaches. A random forest predictive model is trained, and features are ranked based on their contribution to the predictive model. To generate potential and uncertainty maps, compositional data are simulated at unsampled locations via a chain of transformations (isometric log-ratio transformation followed by the flow anamorphosis) and geostatistical simulation. The simulated results are subsequently back-transformed to the original compositional space. The trained predictive model is used to estimate the probability of classes for simulated compositions. The proposed approach is illustrated through two case studies. In the first case study, the major crustal blocks of the Australian continent are predicted from the surface regolith geochemistry of the National Geochemical Survey of Australia project. The aim of the second case study is to discover the superficial deposits (peat) from the regional-scale soil geochemical data of the Tellus Project. The accuracy of the results in these two case studies confirms the usefulness of the proposed method for geological class prediction and geological process discovery. [ABSTRACT FROM AUTHOR]

Big data analytics brings a novel way for identifying geochemical anomalies in mineral exploration because it involves processing of the whole geochemical dataset to reveal statistical correlations between geochemical patterns and known mineralization. Traditional methods of processing exploration geochemical data mainly involve the identification of positive geochemical anomalies related to mineralization, but ignore negative geochemical anomalies. Therefore, the identified geochemical anomalies do not completely reflect the sought geochemical signature of mineralization, leading to uncertainty in geochemical prospecting. In this study, data for 39 geochemical variables from a regional stream sediment geochemical survey of southwest Fujian Province of China were subjected to big data analytics for identifying geochemical anomalies related to skarn-type Fe polymetallic mineralization through deep autoencoder network. The receiver operating characteristic (ROC) and areas under curve (AUC) were applied to evaluate the performance of big data analytics. The AUC of the anomaly map obtained using all the geochemical variables is larger than the AUC of the anomaly map obtained using only five selected elements known to be associated with the mineralization (i.e., FeO, Cu, Pb, Zn, Mn). This indicates that big data analytics, with the support of machine learning methods, is a powerful tool for identifying multivariate geochemical anomalies related to mineralization. [ABSTRACT FROM AUTHOR]

Geochemical stream sediments survey was conducted in the northwestern part of Wadi Allaqi area, Eastern Desert, Egypt. The area comprises Precambrian metasediments, intermediate metavolcanics, gabbro, and serpentinites, with intrusive masses of granites and quartz-porphyry and invaded by several quartz veins. The −1.0-mm size fraction is analyzed for As, Cu, S, Mo, Pb, Zn, Co, Ni, Rb, Ba, Sr, Nb, V, U, Th, Cr, Zr, La, Ce, Nd, and Y. The geochemical survey is supported by heavy minerals study in the −0.125 + 0.0625-mm fraction. The geochemical data were statistically investigated using Q-mode cluster and R-mode factor analyses as well as the enrichment factor. Factors 1 (Zr, Nb, Nd, La, and Y), 2 (V, Sr, and Zn), and 4 (Ba and Rb) are mainly controlled by the lithological characters of the rocks hosting Au-sulfide mineralizations and their accompanied hydrothermal alteration zones. In the mineralization Factor 3 (Cu, S, As, Ce, and Mo), arsenic, Cu, S, and Mo are direct indicators, while Ce is indirect one for the Au-sulfide mineralizations. The Cu-S-As-Mo association with Pb and Zn anomalies in the stream sediments draining the quartz-porphyry point to its porphyry copper mineralization. Cobalt and Ni (Factor 5) are pathfinders for the Fe- and Cu-sulfides, whereas Zn and Pb of Factor 8 are additional pathfinders for the Au-sulfide mineralizations. The southern stream sediments having high U/Th ratios with U-Mo association and draining granites traversed by pegmatites, as well as the stream sediments draining Um Garayat area and the quartz-porphyry stock with high abundance of monazite, zircon, epidote, sphene, and ilmenite, could signify sources of U and Th (Factor 7). Two watershed areas have distinct enrichment factors for arsenic suggesting unexplored extensions of Au-sulfide mineralization linked to the Allaqi shear-zone. The enrichment of the mineralization Factor 3 in the drainage system is mainly controlled by the prevailed mechanical dispersion for the hosting heavy minerals in such arid region with minor role of hydromorphic dispersion. The chemistry and mineralogy of the stream sediments are evidently allied to the drained bedrocks and their hosted mineralizations that signify a promising area for detailed exploration. [ABSTRACT FROM AUTHOR]