Prospectivity Mapping Research Articles

Projection Pursuit Random Forest for Mineral Prospectivity Mapping

Projection Pursuit Random Forest for Mineral Prospectivity Mapping

Mapping and litho‐geophysical classification of potential mining areas by spatial remote sensing and helicopter‐borne gamma‐ray spectrometry (Tikirt, Anti‐Atlas, Morocco)

AbstractThe primary focus of this study is to explore the potential mineral areas. Advanced Spaceborne Thermal Emission and Reflection Radiometer images and radiometric data were combined in the Tikirt region to map hydrothermal alteration zones associated with mineralized deposits. Advanced Spaceborne Thermal Emission and Reflection Radiometer images were analysed using band ratios, principal component analysis and a fuzzy logic model to discover argillic, phyllic, propylitic and iron oxide alterations and to generate a mineral prospectivity map. In addition, mono‐element maps of radiometric elements (K in %, eU in ppm, eTh in ppm) and their behaviour in the ternary image have been elaborated to determine the concentrations of radiometric elements and the variation of radiometric character along the exposed terrains. The F parameter was calculated to target high potassium concentration areas associated with hydrothermal alteration zones. Combining the two methods highlighted four hydrothermal alteration zones considered very promising from a mining point of view. These zones are generally linked to the magma bodies affected by normal faults and exhibit a clear spatial correlation with mineral occurrences.

Evaluating the Performance of Machine Learning and Deep Learning Techniques to HyMap Imagery for Lithological Mapping in a Semi-Arid Region: Case Study from Western Anti-Atlas, Morocco

Accurate lithological mapping is a crucial juncture for geological studies and mineral exploration. Hyperspectral data provide the opportunity to extract detailed information about the geology and mineralogy of the Earth’s surface. Machine learning (ML) and deep learning (DL) techniques provide an accurate and effective mapping of various types of lithologies in arid and semi-arid regions. This article discusses the use of machine learning algorithms, specifically Support Vector Machines (SVM), one-dimensional Convolutional Neural Network (1D-CNN), random forest (RF), and k-nearest neighbor (KNN), for lithological mapping in a complex area with strong hydrothermal alteration. The study evaluates the performance of the four algorithms in three different zones in the Ameln valley shear zone (AVSZ) area at eastern Kerdous inlier, Moroccan western Anti-Atlas. The results demonstrated that 1D-CNN achieved the best classification results for most lithological units. Additionally, the LK-SVM demonstrated good mapping results compared to the other SVM models, as well as RF and KNN. Our study concludes that the combination of the CNN and HyMap data can provide the most accurate lithologic mapping for the three selected region, with an overall accuracy of ~95%. However, this study highlights the challenges in identifying different lithological units using remotely sensed data due to spectrum similarities induced by similar chemical and mineralogical compositions. This study emphasizes the importance of carefully considering and evaluating ML and DL methods for lithological mapping studies, then recommends the high-resolution hyperspectral data and DL models for accurate results. The implications of this study would be fascinating to exploration geologists for Mineral Prospectivity Mapping (MPM), especially in selecting the most appropriate techniques for highly accurate mineral mapping in metallogenic provinces.

Supervised Mineral Prospectivity Mapping via Class-Balanced Focal Loss Function on Imbalanced Geoscience Datasets

Supervised Mineral Prospectivity Mapping via Class-Balanced Focal Loss Function on Imbalanced Geoscience Datasets

Hybridizing K-means clustering algorithm with harmony search and artificial bee colony optimizers for intelligence mineral prospectivity mapping

Hybridizing K-means clustering algorithm with harmony search and artificial bee colony optimizers for intelligence mineral prospectivity mapping

Prospectivity Mapping of Tungsten Mineralization in Southern Jiangxi Province Using Few-Shot Learning

The development of mineral prospectivity mapping (MPM), which aims to outline and prioritize mineral exploration targets, has been spurred by advances in data-driven machine learning algorithms. Supervised data-driven MPM is a typical few-shot task, suffering from a scarcity of labeled data, the over-fitting of models and an uncertainty of predictions. The main objective of this contribution is to propose a robust framework of few-shot learning (FSL), combining data augmentation and transfer learning to enable the generation of prospectivity models with excellent predictive efficiency and low uncertainty. The mineral systems approach was used to transfer a conceptual mineral system into mappable exploration criteria. Synthetic minority over-sampling technique (SMOTE) was employed to augment and balance the labeled dataset, allowing for model pre-training with the large synthetic training dataset of a source domain. The knowledge derived from pre-trained models was then transferred to the target domain by fine-tuning, and the prospectivity model was generated in light of over-fitting and uncertainty assessments. The proposed FSL framework was applied to tungsten prospectivity mapping in southern Jiangxi Province. The results indicated that the SMOTE-ed balanced dataset boosted the classification accuracy in the training process. The FSL models yielded an arch-shaped prediction point pattern which was favorable for focusing potential targets with high probability and low uncertainty. The FSL models achieved a high predictive performance (test AUC = 0.9172) and the lowest quantitative over-fitting value compared to the models derived from the benchmark algorithms of random forest and support vector machine. Four levels of potential targeting zones, considering both predictive efficiency and uncertainty, were extracted from the resulting FSL prospectivity map. The final high-potential and low-risk exploration targets only cover 4.27% of the area, but capture 41.53% of known tungsten deposits, thus achieving a superior predictive performance. This study highlights the capability of FSL framework to control over-fitting and generate high-confidence exploration targets with low levels of uncertainty.

Application of Logistic Regression and Weights of Evidence Methods for Mapping Volcanic-Type Uranium Prospectivity

Pucheng district is a part of the Wuyi Mountain polymetallic metallogenic belt, which is constituted by Archean-Proterozoic metamorphic basements and Mesozoic volcanic-sedimentary covers. Uranium deposits are formed as volcanic-hosted and structural controls. In this study, the hybrid data-driven methods of logistic regression (LR) and weights of evidence (WofE) were applied for the mineral potential mapping of uranium in the Pucheng district. Evidential layers such as volcanic stratum, structure, igneous rock, alteration and radioactive anomaly were used in the mineral prospectivity analyses. The results show that the data-driven methods can not only measure the relative importance of each type of geological feature in uranium controls but also delineate prospective grounds for uranium exploration. The receiver operating characteristics (ROC) curve and under the ROC curve (AUC) were applied to measure the performance of the prospectivity models. The data-driven models are highly capable of mapping uranium prospectivity because AUC is close to 1. The results show that more than 90% of the known uranium deposits occur in regions with high probability. LR performs a little better than WofE in this area. The prospectivity mapping confirmed that there is significant potential for uranium mineralization for further exploration.

Deep Reinforcement Learning for Mineral Prospectivity Mapping

Deep Reinforcement Learning for Mineral Prospectivity Mapping

Search for promising locations to explore geothermal resources based on joint analysis of sections of petro- and thermophysical properties of rocks

The purpose of the study is identification of locations promising for geothermal resource exploration based on the joint analysis of sections of petro- and thermophysical properties of rocks. Having conducted magnetotelluric sounding in the geothermal area under investigation and having built 2-D models of electrical resistivity, porosity, permeability, temperature, thermal conductivity and specific heat capacity, the authors performed a joint cluster analysis of sections of enumerated parameters. A cluster cross-section is constructed taking into account all the discussed parameters, which is essentially a petro- and thermophysical “passport” of the site under examination. Each cluster of the latter is characterized by its own set of property ranges. Two promising locations for drilling exploratory boreholes have been identified in the Soultz-sous-Forêts (France) geothermal area. One of them corresponds to the already developed reservoir of petrothermal energy, while another is located at the depths of 2–3 km in another part of the section and has a potential for future exploration works. The information base created on this basis is a convenient tool for interactive selection of sites promising for exploration of different deposits and building of a “prospectivity map” of surface sites for drilling exploration boreholes.

Geological Controls and Prospectivity Mapping for Manganese Ore Deposits Using Predictive Modeling Comparison: An Integration of Outcrop and Remote Sensing Data, Sinai Microplate, Egypt

Geological Controls and Prospectivity Mapping for Manganese Ore Deposits Using Predictive Modeling Comparison: An Integration of Outcrop and Remote Sensing Data, Sinai Microplate, Egypt

Application of data-driven multi-index overlay and BWM-MOORA MCDM methods in mineral prospectivity mapping of porphyry Cu mineralization

The present study applies proposed hybrid mathematical frameworks, namely Data-driven Multi-Index Overlay Method (DMIO) and Best Worst Method-multi objective optimization by ratio analysis (BWM-MOORA) techniques, for generating robust predictive models enhancing mineral prospectivity analysis is demonstrated. These are used for (i) weighting multiple groups of evidence maps acquired from geological databases, such as remote sensing, geological, geochemical, and geophysical, and then (ii) integrating these weighted layers into robust predictive models to assess porphyry copper prospectivity. This case study of the Chahargonbad area is a district of the volcano-plutonic Urumieh–Dokhtar magmatic arc belt. Porphyry copper deposit (exploration and development) models were used for distilling geoscientific datasets that are available into a group of evidence-based layers. The BWM-MOORA method combines two separate processes in which BWM is applied to weight each criterion and sub-criteria, whereas the MOORA approach helps rank these attributes. Each layer is first converted to another space of [0,1] by the sigmoidal logistic function to perform the critical process of integrating each of the evidence layers. Then, a concentration-area fractal model was used to classify evidences into some geospatial populations. In addition, a prospectivity / mineral potential mapping (MPM) analysis is performed using the DMIO, a well-proven, and accurate data-based methodology, to verify the satisfactoriness and validation of the applied methodology. Prediction-area (P-A) and normalized density (Nd) analysis (plots) were performed as validation methods to estimate the reliability and evaluate the relative efficiency of the MPM models. Therefore, these detailed outcomes confirm the validity of the BWM-MOORA model (Nd = 3.17), which is slightly higher than the valid and proper model DMIO (Nd = 3) in determining the specific potential mineralization regions.

Earth observation approach for targeting stratiform deposit of manganese in central India

The stratiform manganese deposits, hosted within the Sausar Group of rocks, in Madhya Pradesh state of India, has undergone polyphase deformation and metamorphism. The region is an enterprising geosite to understand the lithological and tectonic control on manganese mineralization. We have carried out a conjugate remote sensing based analysis using Visible, Near Infrared and Short Wave Infrared bands (VIS-NIR-SWIR) of optical imagery from multiple sensors supplemented with analysis of L-band microwave data over the Balaghat district of Madhya Pradesh to delineate prospective areas for manganese exploration, supported by the lab and field-based evidence.The enhanced image products (i.e.band ratios, principal components) and spectral anomaly map derived from Advanced Space-borne Thermal Emission and Reflection Radiometer (ASTER) and Sentinel-2B multispectral bands are used to delineate the lithological contrast of the host rock and indicate key locales of mineralization. The geological structures interpreted using the merged product of Advanced Land Observation Satellite Phased Array L-band SAR (ALOS PALSAR −1), Landsat 8 PC bands (2 & 3) and the processed Horizontal Gradient Magnitude (HGM) from advanced satellite gravity model (EIGEN-6C4), illustrates the structural control on the mineralization. The petrological details derived from the host rock samples shows microstructures like micro-folding, foliations, quartz grain elongation etc., thus representing the microscale deformations within the rock units occurring in the area.The lab-based spectral signatures of manganese-bearing host rock was resampled to image spectra and spectral anomaly map using ConstrainedEnergyMinimization(CEM), which shows good correlation with the existing mine locations.The information from remote sensing and lab analysis has been integrated thematically using the knowledge-based Weighted Sum Method (WSM) with existing mine locations as the training dataset. The derived mineral prospect map indicates several new areas of manganese prospectivity which can serve as a guide for detailed future explorations in the region .

Mineral prospectivity mapping for forecasting gold deposits in the Central Kolyma region (Magadan region, Russia)

A geostatistical predictive and prospecting model of gold mineralization (map of potential resource prospects) for the area of the Central Kolyma region (within the Inyali-Debinsky synclinorium) was built using machine learning methods. Multispectral satellite images, materials of medium- and small-scale geological mapping, data of medium-scale geophysical and lithochemical surveys were used as source data. Based on this data, an exhaustive set of indicators related to the search criteria was calculated. To build a forecast, the Random Forest method was used using random negative samples and averaging the result. The obtained forecast maps allow localizing potential gold ore objects with a spatial resolution of 2 km, which makes it possible to select promising areas for licensing. In addition, the possibilities of using machine learning methods for metallogenic zoning of the territory are shown.

Project-scale prospectivity analysis for unconformity-related uranium deposits in the Badami Group of the western Kaladgi basin, India

Developing successful exploration strategies for finding new ore deposits is a major challenge for exploration geologists, particularly, for concealed deposits in terranes with thick sediment cover. Proterozoic unconformity-related uranium deposits occur at, below, or above, unconformable contacts between Archean and Paleoproterozoic crystalline basement and siliciclastic sediments of middle or later Proterozoic age, often at locations where fault systems cut across either one or both of the geologic units that lie above and below the unconformity. The present study implements project-scale prospectivity analyses to identify drilling targets in the Badami Group of the western Kaladgi basin, India. Topographic profiles; field observations of breccia and lineaments; surface and sub-surface sampling; petrography and geochemistry; drill-core lithologs; and interpreted aeromagnetic anomalies were used to build a generalized local geological model of unconformity-related uranium deposits in the study area. Based on the geological model and a generalised mineral systems model, hydrothermal alteration, proximity to lineaments, proximity to unconformity surface, and anomalous uranium content on the surface and in the groundwater were identified as key project-scale targeting criteria. Seven high-priority drilling targets were mapped based on expert driven prospectivity analysis constrained by a GIS-based fuzzy prospectivity modelling. Out of these, four target areas coincide with the targets identified by the Atomic Minerals Directorate (AMD) of India based on their independent investigation. The three new targets from this study are recommended as targets for detailed exploration and drilling. In addition to this, about eight additional low priority targets were identified for follow-up exploration activities. This paper demonstrates that the prospectivity mapping methods can be applied at the project scale. The strategy demonstrated in this study can be used for exploration targeting in similar basins like the western Kaladgi basin that are structurally less disturbed.

Mineral prospectivity mapping using attention-based convolutional neural network

Data-driven mineral prospectivity mapping (MPM) based on deep learning methods has become a powerful tool for mineral exploration targeting in the past years. Convolutional neural networks (CNNs) have shown great success in this field because of their powerful ability to capture the complex spatial geo-anomalies related to mineralization. However, the exploration big data applied to MPM mainly relies on the high dimensions of evidence layers (other than spatial dimensions), namely, a large number of channels. This impedes the extraction of key channel features related to mineralization when using traditional CNNs. In this paper, we developed an ensemble MPM method based on CNN and Attention model: the ATT–CNN method. Specifically, a channel attention layer is added after the convolution operation of the CNN to enhance the extraction of key channel features in complex exploration data, thereby improving the feature extraction ability and prediction accuracy of CNN for MPM. A case study of W–Sn mineral prospectivity modeling in the Nanling metallogenic belt in South China was used to verify the proposed method. To alleviate the issue of training sample scarcity, we used data augmentation methods (including sliding window and random zero noise addition) when training CNN models. The results show that the prediction accuracies of the ATT–CNN model (92.949% and 94.872% using sliding window and random zero noise addition, respectively) are higher than those of the traditional CNN (91.667% and 92.308%, respectively). Moreover, the improved areas under the receiver operating characteristic curves (AUC) of ATT–CNN (0.987 and 0.971) compared to those of the CNN (0.970 and 0.964) suggest that the proposed ensemble method improves the geological generalization of CNN. The high agreement with known deposits suggests that the areas targeted in this study can guide future mineral exploration of the W–Sn mineralization in the Nanling range.

Data-driven logistic function for weighting of geophysical evidence layers in mineral prospectivity mapping

Logistic function has been extremely executed to transform values of evidence layers into the range [0,1] to produce evidence maps assimilated or to predict mineralization zones as fuzzy prospectivity models. Many researchers have recently presented fuzzy layers transformed with continuous weights in range from [0, 1]. But thay did not try to consider efficiency or inefficiency of achieved results by overcoming exploratory bias. However, majority of integration methods comprising an inherent bias due to employing expert's judgments or applying trial-and-error procedure for determining some parameters of their functions. This research demonstrated, the application of data-driven fuzzy logic to assign continuous weights to geophysical evidence maps which can be integrated to generate prospectivity map with no biasness (with a 91% prediction ability). Results of the fuzzy transformation as continuous weights and discrete weights were compared and validated using prediction-area plot and locations of some mineralization clues.

Territory Suitability Assessment for Conducting Detailed Geological and Mineralogical Mapping Based on Statistical Methods of Remote Sensing Data Processing Landsat-8: A Case Study in the Southeastern Transbaikalia, Russia

The work considers the suitability of using multispectral satellite remote sensing data Landsat-8 for conducting regional geological and mineralogical mapping of the territory of south-eastern Transbaikalia (Russia) in conditions of medium- low-mountain relief and continental climate. The territory was chosen as the object of study due to its diverse metallogenic specialization (Au, U, Mo, Pb-Zn, Sn, W, Ta, Nb, Li, fluorite). Diversity in composition and age of ore-bearing massifs of intrusive, volcanogenic and sedimentary rocks are also of interest. Statistical processing algorithms to increase spectral information content of satellite data Landsat-8 were used; they include: principal component analysis (PCA); minimum noise fraction (MNF) and independent component analysis (ICA). Eigenvector matrices analysed on the basis of statistical processing results and two-dimensional correlation graphs were built to compare thematic layers with geological material classes: oxide/hydroxide group minerals containing transition iron ions (Fe3+ and Fe3+/Fe2+); a group of clay minerals containing A1–OH and Fe, Mg–OH; minerals containing Fe2+ and vegetation cover. Pseudo-coloured RGB composites representing the distribution and multiplication of geological materials classes was generated and interpreted. Integration of informative thematic layers with using fuzzy logic model was carried out to construct a prospectivity map. Received map was compared with geological information, and positive conclusions about territory suitability for further remote mapping research of hydrothermally altered zones and hypergenesis products in order to localize areas promising for identifying hydrothermal-metasomatic mineralization were made.

Regional Quantitative Mineral Prospectivity Mapping of W, Sn, and Nb-Ta Based on Integrated Information in Rwanda, Central Africa

As the need to discovers new mineral deposits and occurrences has intensified in recent years, it has become increasingly apparent that we need to map potentials via integrated information on the basis of metallogeny. Occurrences of mineralization such as tungsten (W), tin (Sn), columbium (Nb), tantalum (Ta), gold (Au), copper (Cu), lead (Pb), zinc (Zn), manganese (Mn) and monazite (Mnz) have been discovered in Rwanda. The objective of this study was to present a regional quantitative mineral prospectivity mapping (MPM) of W, Sn and Nb-Ta mineralization in Rwanda using the random forest (RF) method on the basis of open source data, such as geological maps, Bouguer gravity anomalies, magnetic anomalies, Landsat 8 images, ASTER GDEM, Globeland30, and OpenStreetMap. In addition, a newly introduced interpolation–density–delineation (IDD) process was applied to deal with the blank (masked) areas in remotely sensed mineral alteration extraction. Additionally, a k2-fold cross-validation method was also proposed to obtain more reasonable test errors. Firstly, the metallogenic regularity of W, Sn and Nb-Ta in Rwanda was summarized with the help of articles online. Secondly, original geological, geophysical, and remote sensing data were utilized to generate secondary data. Specifically, the IDD process was applied subsequent to the directed principal component analysis method (DPCA) to reconstruct the alteration anomaly map, and a relevant dataset was formed by the combination of original and secondary data. Thirdly, specific predictor layers for W, Sn and Nb-Ta were selected from relevant data via spatial correlation with known deposits, respectively, and the predictive models were established. Finally, near 26,000 squares were zoned in Rwanda, and RF was optimized and applied, the k2-fold cross-validation method was utilized to assess test errors, metallogenic belts and prospective areas for W, Sn, and Nb-Ta were delineated on the basis of total mineralization potential map and likelihoods map. Results proved that the open source data online were valid for drawing a preliminary mineralization potential map. Furthermore, it was also shown that the IDD method is suitable for the postprocessing of masked alteration anomaly maps. Belt IV-4 in the northwest and belt IV-2, IV-1 in the middle-east of Rwanda, containing a number of prospective areas, possess considerable likelihoods of deposits, and mining in Rwanda is at its dawn, with potential worth expecting.

Geochemical multifractal modeling of soil and stream sediment data applied to gold prospectivity mapping of the Pitangui Greenstone Belt, northwest of Quadrilátero Ferrífero, Brazil

Geochemical mapping has become essential to prospective assessments in both brownfield and greenfield exploration targeting. During the last two decades, several techniques have been applied to the identification of mineralization's geochemical signature. Among these techniques, Hierarchical Clustering (HC) and Principal Component Analysis (PCA) have been widely used. Other techniques, like Multifractal Spectrum (MS) and Local Singularity Mapping (LSM), are efficient in the generation of models suitable to determine the type of database statistical distribution and mapping zones of geochemical enrichment and depletion. We combined these techniques for soil and stream sediment geochemistry and generated two prospective geochemical models for the gold deposits from the Pitangui Greenstone Belt (PGB), located at northwest of Quadrilátero Ferrífero – Brazil. The results of PCA and HC showed a geochemical association of As-Bi-Mo-Sb with gold mineralization in the PGB. According with descriptive statistical analysis, these elements do not display normal or log-normal statistical distribution, which allowed the application of multifractal models on the geochemical mapping. MS technique revealed that all elements associated with gold have a multifractal distribution. In addition, geochemistry data shows that high contents of elements measured in the stream sediment analysis displays greater dispersion when compared to elements of soil geochemistry data, due to their mobility in environmental conditions. We used LSM technique to map geochemical enrichment of the elements associated with gold. The maps were evaluated using the Ground Truth Analysis (GTA) and the results showed a good performance in the identification of prospective zones. The anomalies identified in both geochemical maps showed a strong correlation between geochemical enrichment zones and geological factors that favors gold mineralization. These results bring a new perspective on the distribution of geochemical anomalies using, for the first time, the multifractal geochemical modeling on the PGB. The geochemical maps point to new targets for mineral exploration and contribute significantly for the prospective models of the region.

3D modeling of detachment faults in the Jiaodong gold province, eastern China: A Bayesian inference perspective and its exploration implications

The world-class Jiaodong gold province, which is located in the Jiaodong Peninsula, eastern North China Craton, hosts > 5,000 t of gold resources. The major gold mineralization in this area is mainly controlled by NE-NNE faults, such as the Zhaoping, Jiaojia, and Sanshandao faults. While intensive efforts have been devoted to the investigation of the architecture of the three detachment faults, the fault geometry at depth remains highly uncertain and hotly controversial due to the scarcity of observations, non-uniqueness of the geophysical inversion, and underutilization of the geological knowledge. Along these lines, the deep three-dimensional (3D) geometry of detachment faults from a Bayesian inference perspective is reconstructed in this work. In the Bayesian framework, the multi-source information and the prior geological knowledge are combined into an integrated probabilistic model and the posterior probability distribution of detachment faults is inferred. The resulting posterior distribution permits an estimation of the uncertainty (information entropy) of the 3D fault geometry and the expected shape undulations compared with a prior reference model. More specifically, it demonstrates that specific and detailed 3D models of faults could be inferred by tracking surfaces with maximum information entropy of the geological units. The reconstructed models exhibit detailed 3D geometry of the fault surfaces and their reliability is validated by a recent drill hole at ∼−2,300 m depth. In terms of the reconstructed 3D models, the dip variations are analyzed and the possible stress directions at deep zones of the three faults are discussed. The shape features of the reconstructed faults are also quantified, which indicates that the fault sections with a steep-to-gentle transition in dip angle are closely associated with the gold mineralization. The 3D gold prospectivity mapping with the spatial features as predictor variables, highlights seven deep prospecting targets, which are generally consistent with the gently dipping zones and beneath the known orebodies.