Uncertainty Quantification in Mineral Resource Estimation

Estimations of mineral resources and ore reserves have been recently widely used by mining engineers and investors. The classification framework based on the prepared code by the Joint Ore Reserves Committee of The Australasian Institute of Mining and Metallurgy, Australian Institute of Geoscientists and Minerals Council of Australia (JORC code), which is one of the international standards for mineral resource and ore reserve reporting, provides a template system that conforms to international society requirements. Recent research has shown that an existing fault risk can affect the mineral resource and ore reserve estimation. According to this research, the faulted area that is involved in the effect on the estimated region is so extensive that it is not distinguishable. In this research, a new method called FGT (F for fault, G for grade and T for thickness) is introduced and presented for the estimation of mineral resources. The proposed method can provide an error map of a particular aspect of the combination of coal accumulation (G), fault risk (F) and thickness (T), and its output would categorise the mineral resources. This method was implemented in the Parvadeh Ш coal deposit, which is located in the eastern portion of Central Iran. The deposit contains five seams named B1, B2, C1, C2 and D; of these, C1 was selected as the most important seam in the exploratory grid analysis. Thus, C1 alone can reflect the properties of the Parvadeh Ш deposit. In this study, we compared the conventional method and the FGT method. This comparison indicated that the areas that should be rejected from the region in the FGT method are less and more distinguishable than those determined with the conventional method. Therefore, the inferred resources can be completely differentiated from the indicated and measured resources with a high resolution. The conventional method cannot distinguish between these three categories at this level of resolution. Therefore, the FGT approach has high precision in classifying the coal resource compared to the conventional method.

A resource estimate (calculation of total tonnes and grade within a deposit) is generally underpinned by a clear understanding of the geometry, orientation and control of mineralization and set of assays with assigned 3D spatial locations. The proper interpretation and handling of outliers in a data set is crucial to estimating a mineral resource that is representative of the deposit. Without a firm understanding of the distribution of metal in a deposit, mine planning, scheduling, process planning and economic analysis will likely be flawed. Three of the most important issues in the mineral resource estimation process are the recognition of outlier values in a data set, the source of the outlier values, and the subsequent handling of these high values. Treatment of outliers in mineral resource estimations is a perplexing problem for which there is no generally accepted solution. Each deposit may have a unique distribution of outlier values which may require multiple methods of treatment to fully understand the issues and their effect on the mineral resource estimate.

Accurate prediction of mineral grades is a fundamental step in mineral exploration and resource estimation, which plays a significant role in the economic evaluation of mining projects. Currently available methods are based either on geometrical approaches or geostatistical techniques that often considers the grade as a regionalised variable. In this paper, we propose a grade estimation technique that combines multilayer feed-forward neural network (NN) and k-nearest neighbour (kNN) models to estimate the grade distribution within a mineral deposit. The models were created by using the available geological information (lithology and alteration) as well as sample locations (easting, northing, and altitude) obtained from the drill hole data. The proposed approach explicitly maintains pattern recognition over the geological features and the chemical composition (mineral grade) of the data. Prior to the estimation of grades, rock types and alterations were predicted at unsampled locations using the kNN algorithm. The presented case study demonstrates that the proposed approach can predict the grades on a test dataset with a mean absolute error (MAE) of 0.507 and R2=0.528, whereas the traditional model, which only uses the coordinates of sample points as an input, yielded an MAE value of 0.862 and R2=0.112. The proposed approach is promising and could be an alternative way to estimates grades in a similar modelling tasks.

Rock density information is needed for the estimation of ore tonnage, and also for determining the amount of contained metal; this is because mineralisation grades are usually reported as a ratio of the economically viable components, metals or minerals, to weight units (e.g. grams/tonne). Despite the obvious importance of rock density for the accurate estimation of tonnage and grade of mineral resources, this parameter is often overlooked and receives significantly less attention than assayed metal grades. In particular, the quantity and spatial distribution of the rock density measurements may be chosen without considering the impact of the rock density measurements on the mineral resource estimation. This paper attempts to address several issues associated with current practices of rock density modelling. The first part of the paper overviews the methods most commonly used for measuring the dry bulk density (DBD) of rocks for the estimation of mineral resources. This is followed by a second part proposing the methodology of geostatistical modelling of the rock densities which is suggested as a mathematically supported approach for choosing optimal sampling grids for density samples. It is supported by several case studies of the deposits where geostatistically optimal DBD sampling grids were estimated and proposed for definition of the Measured and Indicated resources and for the grade control purpose. It was noted that the exploration team needs to assure that DBD data collected during the drilling campaigns are not only sufficient for definition of resources and reserves but are also closely enough spaced to allow accurate grade control. The optimal for grade control DBD grids, according to the geostatistical analysis, is similar to the chemical assay grids required for definition of the Measured resources or, less commonly, Indicated resources of the studied deposits.

Radial basis functions can be used as an alternative method for mineral resource estimation when ordinary kriging cannot be done because it is impossible to calculate an experimental variogram. Interpolation based on radial based functions is a method that is similar to ordinary kriging. In this context, this article presents the results of gold resource estimation and modeling of the União Luis and Morro do Carrapato gold deposits located in Alta Floresta Gold Province, Mato Grosso. Experimental variograms in the main direction (90-degree strike) resulted in a pure nugget effect, whereas in the orthogonal direction (180/85), they resulted in a structured variogram. However, the structured variogram for 180/85 is useless because three orthogonal variograms are needed to define the anisotropy ellipsoid for deriving a correlation model used in ordinary kriging. Thus, mineral resource estimation was done using multiquadric equations, a very popular radial based function kernel. The 3D model for gold grades showed a vertical distribution suggesting a structural conditioning for gold mineralization. The grade-tonnage curve with a simulated cutoff of 5.0 g/ton resulted in a gold mineral resource of 2.122 tons. Through the above, the gold estimation and multiquadric equation-based 3D model for the studied area can be considered effective in its objective to estimate ore resources from the available data.

Both uncertainty assessment and validation have shown that the current global products of phytoplankton functional types (PFT) on Copernicus Marine Service for the Arctic Ocean (AO) bear larger gaps and higher uncertainties compared to that in the low latitude oceans. In the framework of Copernicus Marine Service Evolution Program, we propose a project ML-PhyTAO to exploit marine big data-driven machine learning (ML) methods in the PFT monitoring for high latitudes, and aim to set up an improved algorithm for better quantifications of multiple PFTs in the AO. A large marine data set (including bio-optical, biogeochemical, and physical data) obtained from various sources will be exploited as inputs for algorithm training and validation. The ML-PhyTAO is expected to deliver improved gap-free products of several key PFTs (diatoms, haptophytes, dinoflagellates, chlorophytes and prokaryotes) with uncertainty assessment to complement the current ocean colour/ biogeochemical data sets for the AO on the Copernicus Marine Service Data Store. Such PFT data set with improved accuracy will allow reliable long-term monitoring and trend analyses for the surface phytoplankton community structure, helping in detecting potential shifts and changes in phytoplankton diversity in the AO under the Arctic amplification effect. In this work we will demonstrate the framework of the project and present our latest outcome from the project by showing our first results on the experiments of ML methods using our well compiled training data sets including the in situ PFT data, satellite and model simulated data/products from Copernicus Marine Service covering various optical/physical/biogeochemical parameters.

A dificuldade na quantificação do grau de incerteza associado com a estimativa de recursos minerais tem levado ao desenvolvimento de uma gama de metodologias, termos e definições. Isso gerou uma situação em que praticamente cada companhia de mineração possui sua metodologia particular de classificação. Intuitivamente, sabe-se que os métodos tradicionais utilizados na avaliação de recursos não levam em consideração a continuidade espacial dos teores e, mesmo que algumas metodologias baseadas em métodos geoestatísticos são incapazes de fornecer uma medida do erro associado às estimativas. Nesse sentido, esse artigo se propõe a desenvolver uma metodologia que permita uma análise quantitativa e qualitativa dos recursos minerais estimados por meio da incorporação da incerteza e da adequada definição do risco ou erro associado às estimativas. Um estudo de caso com um banco de dados de minério de ferro típico foi efetuado, permitindo uma comparação, tanto entre os diferentes sistemas de classificação, como entre os parâmetros-chave de cada sistema, sendo que os resultados obtidos permitiram ressaltar as limitações específicas de cada metodologia, além de demonstra a natureza empírica dos métodos tradicionais, devido ao seu caráter predominantemente subjetivo.

Bulk density miscalculations and subsequent reporting can produce resource estimation errors and unexpected financial outcomes. Density and bulk density vary throughout a mineral deposit, especially in porous deposits and deposits containing variable amounts of high-density minerals (e.g., barite). Resource estimators must correlate bulk density with core recovery and apply quality assurance/quality control analysis to bulk density data. Core recovery is an important factor in resource estimation and reporting. Block tonnages (ore and waste) are estimated from volumes using a dry bulk density value. Ore dilution is a volumetric effect integral to ore reserve estimation that also has the potential to affect the economics of mining a deposit. Both the bulk density and dilution problems could be overcome by reporting a mineral resource estimate in grade per unit volume.

The basis of accurate mineral resource estimates is to have a geological model which replicates the nature and style of the orebody. Key inputs into the generation of a good geological model are the sample data and mapping information. The Obuasi Mine sample data with a lot of legacy issues were subjected to a robust validation process and integrated with mapping information to generate an accurate geological orebody model for mineral resource estimation in Block 8 Lower. Validation of the sample data focused on replacing missing collar coordinates, missing assays, and correcting magnetic declination that was used to convert the downhole surveys from true to magnetic, fix missing lithology and finally assign confidence numbers to all the sample data. The missing coordinates which were replaced ensured that the sample data plotted at their correct location in space as intended from the planning stage. Magnetic declination data, which was maintained constant throughout all the years even though it changes every year, was also corrected in the validation project. The corrected magnetic declination ensured that the drillholes were plotted on their accurate trajectory as per the planned azimuth and also reflected the true position of the intercepted mineralized fissure(s) which was previously not the case and marked a major blot in the modelling of the Obuasi orebody. The incorporation of mapped data with the validated sample data in the wireframes resulted in a better interpretation of the orebody. The updated mineral resource generated by domaining quartz from the sulphides and compared with the old resource showed that the sulphide tonnes in the old resource estimates were overestimated by 1% and the grade overestimated by 8.5%.

Application of risk assessment to improve sustainability in bauxite mining

The advance of electromobility has boosted global demand and interest in lithium. The consequent expansion of lithium mining puts the sustainability of Chile's Andean salt flats at risk. In these unique ecosystems, lithium is abstracted from mineralized groundwaters, referred to also as brines. This article analyses the legal treatment of brines and its implications in assessing the socioecological impacts of lithium mining projects. For our analysis, we draw from scholarship at the intersection between hydrosocial research and critical legal geography. Methodologically, our study is based on interviews and the analysis of legal texts and judicial and administrative claims, including the environmental impact assessment studies of the three single lithium mining projects approved in the country. We show that the interpretation of brines as mining resources supported by mining companies and endorsed by environmental State agencies is based on a legal loophole. We document how such interpretation is operationalized and contested in the environmental impact assessments of three mining projects and other instances. We explore how the same legal loophole could lead to alternative interpretations and relatedly regulatory proposals and discuss their implication for the assessment of socioecological impacts of mining projects. These include first an understanding of brines as hybrids minerals/waters put forward in a recent report commissioned by State agencies, and second an interpretation of brines as a type of water. The latter is in line with the position of some indigenous groups and academics. We conclude with reflections on the implications of our analysis for lithium mining in Chile and beyond.

Non-parametric machine learning methods for interpolation of spatially varying non-stationary and non-Gaussian geotechnical properties

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.

The subject of the study is the forecast of the development of the Arctic territories of Russia at the zonal, regional and municipal levels based on the concept of technical and economic (structural) dynamics of N. Kondratiev – S. Glazyev – С. Peres. The object of study is the Arctic zone of the Russian Federation. The purpose of the study is to describe the technical and economic dynamics of the Arctic territories until 2035. This goal determined the solution of three tasks: 1) identifying the features of new resource development of the Russian Arctic based on the mineral resources and fuel and energy potential of the Arctic territories of Russia; 2) justification of the current situation in the regions of the Russian Arctic, such as the dynamics of technological structures and calculated positions on the volume of GRP in 2035; 3) compiling a list of Arctic municipalities (49) with limited delivery times for cargo and their typology based on the speed of technical and economic dynamics. Results of the study: I. The resource potential of the Russian Arctic from the point of view of technological structures can be strengthened through the commissioning of deposits of graphite, antimony, and bauxite, the global demand for which is characterized by favorable conditions in the forecast period; II. The regions of the Russian Arctic are differentiated based on the analysis of greenfield projects carried out during the forecast period into three groups of technological renewal: a) maximum; b) partial; c) minimal. The forecast of the GRP of the Arctic territories based on resource production indicators in 2035 determined their rank: 1) Yamal-Nenets Autonomous Okrug; 2) Murmansk region; 3-4) Nenets Autonomous Okrug and the Arctic of the Krasnoyarsk Territory; 5) Arctic of the Republic of Sakha (Yakutia); 6) Arctic of the Komi Republic; 7) Chukotka Autonomous Okrug; 8) Arctic of the Arkhangelsk region; 9) Arctic of the Republic of Karelia. III. Arctic municipalities with limited delivery times are separated into four groups from the point of view of the prospects for the formation of a new technological structure based on new projects: 1) maximum potential (12); 2) average potential (9); 3) without new mining projects, but with prospects for completing the development of previously started ones and technological modernization of old projects (12); 4) without prospects for the deployment of new projects for the development of non-renewable resources, but with prospects for the development of biological resources (16). Possible application of the results can be documents of strategic planning of the Russian Arctic at the zonal, regional and municipal levels.

Mining companies and communities: Collaborative approaches to reduce social risk and advance sustainable development