Results

The Multi-source and Multi-scale Earth observation and Novel Machine Learning Methods for Mineral Exploration and Mine Site Monitoring (MultiMiner) project develops novel data processing algorithms fo…

The Multi-source and Multi-scale Earth observation and Novel Machine Learning Methods for Mineral Exploration and Mine Site Monitoring (MultiMiner) project develops novel data processing algorithms for cost-effective utilization of Earth Observation (EO) technologies for mineral exploration and mine site monitoring. The project focuses on new EO-based exploration technologies for critical raw materials (CRM) to increase the probability of finding new sources within EU thereby strengthening the EU autonomy in the area of raw materials. The importance of good data management practices is emphasized in data-intensive projects such as MultiMiner. During the project, large and heterogenous datasets will be acquired and analyzed. The project comprises data acquired in different scales and by different technologies. The number of the consortium partners (13) highlights the importance of data harmonization, i.e. efforts to collect and distribute data using collectively agreed methods, formats and metadata descriptions. MultiMiner generates and collects a range of data types, most of which are instrument data. Examples of such data include, but are not limited to, point spectral data, radar satellite data and laboratory-collected hyperspectral image data. These data are stored in internationally accepted file formats for their effective sharing, re-use, and preservation. The total estimated size of the data is in the order of terabytes. MultiMiner is reusing publicly available satellite data (e.g. EnMAP, Sentinel) and existing field data from participating test sites. Should any personal data be collected during the project (e.g., stakeholder surveys), all such data are stored in line with the requirements of the General Data Protection Regulation (GDPR). By default, all data, software and algorithms needed to reproduce and facilitate the reuse of the results of MultiMiner are made publicly available using Zenodo and GitHub. These are repositories that provide persistent identifiers (e.g., DOIs), generate formal citations and whose metadata and data are registered and searchable by search engines to ensure the findability of the data. The MultiMiner consortium shares any data as soon as possible and usually immediately after the data production and quality controls. Here, we follow the principle “as open as possible, as closed as necessary” and should the immediate release of the data be delayed, or the data restricted e.g., due to IP considerations, it is always ensured that the metadata of the data are immediately and publicly available. Only such trusted data repositories are used to share data which provide harvestable and public machine-accessible interfaces to metadata, data files, dataset terms, licenses and version information, making the data findable and accessible. When possible, all data are prepared such as to include qualified references to other data, maximizing the interoperability of the data. Data are carefully documented at all project phases (data acquisition, data analysis, data sharing) and metadata are compiled using such broad cross-domain and domain-specific metadata schemas (e.g., DataCite, Dublin Core, ISO 19115) and controlled vocabularies (e.g., ISO 3166) which maximize the reusability and interoperability of the data. The metadata of the data always include a detailed provenance, thus including such items as the names of the authors, in addition to information about how the data was collected and processed. By default, MultiMiner shares all its data using clear and permissive licenses to maximize data reusability. By default, all metadata and data are shared using the Creative Commons Public Domain Dedication CC0 and software, algorithms and apps generated in the project are shared using permissive software licenses such as the MIT license. Access to physical samples is given upon request and at the same time as the associated data, following the same principles of openness. Only if data sharing is against the beneficiary’s legitimate interests, including, but not limited to, commercial exploitation of the research outputs, or if the consortium is reusing data already shared by other organizations (e.g., the satellite data) data is not shared or access to it is restricted. All data is stored in the secure servers of the project partners during the project and a minimum of 10 years after the end of the project. Cloud services that are deemed safe are used to share the data between the said partners. While the Coordinator takes a leading role in data management, it is the responsibility of the individual partner organizations to ensure that the data they generate in the project is managed in accordance with the DMP and meets appropriate national and international legal and ethical standards. Data management activities are adequately resourced and implemented, and all partners allocate a minimum of 5% of their time for research data management. The DMP is treated as a living document and is updated throughout the project, and we publish its different versions in publishing platforms such as Research Ideas and Outcomes RIO journal. Examples of situations where the DMP would be updated include, but are not limited to: new datasets, changes to data storage or data sharing practices, changes in the level of openness of the data (fully open, partially open or fully closed). A new version of the DMP will be created and shared any time major changes take place. 

Magnesit ist ein essentieller Rohstoff für die Feuerfestindustrie und damit von wesentlicher Bedeutung für den Industriestandort Europa. Österreich ist für diesen Rohstoff seit jeh…

Magnesit ist ein essentieller Rohstoff für die Feuerfestindustrie und damit von wesentlicher Bedeutung für den Industriestandort Europa. Österreich ist für diesen Rohstoff seit jeher einer der wichtigsten europäischen Abbau- und Innovationsstandorte und spielt traditionell eine Hauptrolle bei der Erforschung von Spatmagnesit. Bei der Spatmagnesitbildung wird im Wesentlichen ein karbonatisches Ausgangsgestein wie Dolomit durch metasomatische Prozesse in Magnesit umgewandelt. Wie und unter welchen Bedingungen diese Mineralisierung in den Ostalpen stattgefunden hat, ist aber im Detail unklar und trotz der langen Forschungsgeschichte gibt es noch immer keinen Konsens über die Entstehung dieses Magnesittyps in Österreich. Im Rahmen des internationalen Horizon Europe Projekts MultiMiner* und mit finanzieller wie logistischer Unterstützung der RHI Magnesita AG, wird der Magnesit-Bergbaubezirk Hochfilzen in Hinblick auf diese Problematik neu bearbeitet. Neben geologischen Kartierungen durch die GeoSphere Austria und Untersuchungen zur Bildung des Magnesits durch die Montanuniversität Leoben geht es in diesem Projekt insbesondere um die Weiterentwicklung von Fernerkundungsmethoden. Dafür werden nicht nur Drohnen eingesetzt, sondern es werden auch Satellitendaten interpretiert, Hyperspektralmessungen mineralogisch ausgewertet und die Vegetation auf Abraumflächen untersucht. Dabei spielt vor allem die Kalibrierung von Satelliten und „airborne“ Methoden durch geologische Kartierungen sowie durch geochemische und mineralogische Analysen von Gesteinsproben („ground truthing“) eine entscheidende Rolle. Grundlage dieser Methoden ist ein ganzheitliches Verständnis der Spatmagnesitentstehung. Dafür werden unter anderem Ergebnisse der Elektronenstrahlmikrosonde, LA-ICP-MS, µ-RFA und Raman-Spektroskopie herangezogen um so ein modernes Genesemodell in einem größeren geodynamischen Kontext zu finden. Mit der Bearbeitung des Bergbaubezirks Hochfilzen sollen wegweisende Methoden für Prospektionsarbeiten auf Spatmagnesit etabliert werden und die Identifikation von höffigen Gebieten durch neue Genesevorstellungen erleichtert werden.   * This work was funded by the European Union under the Horizon Europe grant No. 101091374 of MultiMiner project - Multi-source and Multi-scale Earth observation and Novel Machine Learning Methods for Mineral Exploration and Mine Site Monitoring.

Dataset description: Monitoring water quality is a critical component of environmental management at mining sites worldwide. This dataset accompanies the manuscript, "Geometry Optimization of Unmanned…

Dataset description: Monitoring water quality is a critical component of environmental management at mining sites worldwide. This dataset accompanies the manuscript, "Geometry Optimization of Unmanned Aerial Vehicle (UAV) Hyperspectral Surveys for Water Quality Retrieval and Satellite Upscaling in Mining Impacted Waters" It contains UAV-based visible and near-infrared (VNIR) spectra, associated metadata, and laboratory water chemistry data used in the study. Data collection and equipment: The dataset features Savitzky–Golay filtered reflectance spectra acquired over mining-impacted waters in the Kassandra and Kirki districts of Greece. The spectral data were collected using an Ocean Optics STS-VIS sensor (337–823 nm; 40 × 42 × 24 mm) mounted on a DJI Phantom 3 Advanced platform. Alongside the spectra, metadata is provided, including acquisition geometry (flight direction relative to water flow and solar azimuth) and site identifiers. For every sampling location, the spectral data are paired with corresponding laboratory measurements of water quality parameters, including total suspended solids (TSS), dissolved iron (diss Fe), and various other comprehensive hydrochemical analytes. Project context: Please note that this dataset includes data for only two of the three study areas discussed in the associated manuscript (Chalkidiki and Kirki). These two specific localities were surveyed as part of the MultiMiner project (European Union’s Horizon Europe research and innovations actions programme under Grant Agreement No. 101091374.). Dataset structure and file overview: Dataset_description.docx: A document providing an overview of the dataset and containing the detailed information presented in this text. Localities_of_water_samples_and_spectral_measurement.csv: A master spreadsheet containing geospatial and descriptive metadata for all sampling locations. It includes longitude, latitude, altitude, and a brief description of the water body type (e.g., stream, pool, lake). The dataset is organized into one summary file and two main directories: In_Situ_Spectrum (Folder): This directory contains 12 files comprising the spectral data from both the Kassandra and Kirki areas. The data within these files are organized by: Filter settings: Processed using three different Savitzky-Golay filter configurations. Flight direction: Categorized by the drone's flight path relative to the target (upstream, downstream, diagonal, as well as a combined/aggregated mean dataset). Lab_Analysis (Folder): This directory contains the results of the comprehensive laboratory water chemistry analyses for all sampled geographic points in Chalkidiki and Kirki. Sun_Position (Folder): This folder contains two files providing time-series sun position metrics for the Chalkidiki and Kirki study areas in Greece. Each record pairs UTC timestamps with solar geometry parameters including solar azimuth, and sun elevation computed for the specific site coordinates.

Carbonate-hosted sparry magnesite is the most important raw material for the production of magnesia sinter (MgO), which is essential for the refractory industry. In one of the most significant ma…

Carbonate-hosted sparry magnesite is the most important raw material for the production of magnesia sinter (MgO), which is essential for the refractory industry. In one of the most significant magnesite mining districts of the Eastern Alps near Hochfilzen (Austria), fine-grained magnesite occurs in two settings: (1) as minable deposits in Silurian to Devonian dolostones and (2) in Permian breccias of red beds in the hanging wall. Despite the far-reaching implications of a possible genetic relationship between these groups on both the timing and the mechanism of sparry magnesite formation in the Eastern Alps, there is no comparative analytical data available. To decipher their relationship, we used electron probe microanalysis, laser ablation inductively coupled plasma mass spectrometry, micro-X-ray fluorescence mapping and Raman spectroscopy.Paragenetic parallels, comparable rare earth element patterns, similar trace element contents and enrichment patterns point towards a common mineralizing fluid of the two magnesite groups. Based on our results, we suggest that a Mg-rich, seawater-derived brine descended into the subsurface, migrated through the red beds, magnesitized the Permian breccias, entered the underlying pre-Permian footwall along fault planes and formed the magnesite deposits – the magnesite groups are therefore related by a common fluid. We contextualize this formation model with the Permian-Triassic metallogenetic epoch in the Eastern Alps and argue for a Triassic ageof the magnesite.

Carbonate-hosted sparry magnesite is the most important raw material for magnesia (MgO) in refractories, which are essential for the metal industry. Several occurrences of this commodity in&…

Carbonate-hosted sparry magnesite is the most important raw material for magnesia (MgO) in refractories, which are essential for the metal industry. Several occurrences of this commodity in the Eastern Alps have been mined and investigated for many decades, but there is still no consensus on their formation. One of them is the Hochfilzen mining district (Tyrol, Austria) where magnesite is associated with Silurian to Devonian dolostones and with Permian breccias in the hanging wall. Up to now, it is unclear if these two magnesite groups are genetically related. To investigate their relationship and improve our understanding of magnesite formation in the Eastern Alps, we used electron microprobe analysis, laser ablation inductively coupled plasma mass spectrometry, micro-X-ray fluorescence mapping and Raman spectroscopy.The deposits of the carbonate-hosted, finely crystalline magnesite display a characteristic succession of metasomatic replacement of dolomite by magnesite and a post-formational retransformation of magnesite to dolomite. The latter stage is also accompanied by the formation of anhydrite, Ca-rich dolomite and calcite. The magnesitized dolomite clasts of the Permian red beds in the hanging wall of the magnesite deposits show similar paragenetic and morphological features as the magnesite mineralization in the mines. However, partial magnesitization of the breccia´s matrix and magnesite rimming dolomite clasts are testimony to the metasomatic in-situformation of magnesite and a rebuttal of magnesite redeposition from the deposits. Additionally, the trace elements in both magnesite groups indicate a common formation merely differing in stratigraphic position. Finally, both magnesite groups were overprinted by regional metamorphism at 286 ± 27 °C and the comparison with coarsely crystalline magnesite from other occurrences in the Eastern Alps shows that metamorphic temperatures do not control the magnesite crystal size. By combining investigations of magnesite in different lithologies and stratigraphic units we present a novel perspective on magnesite formation in the Eastern Alps and a direction forfuture studies. This work was funded by the European Union under the Horizon Europe grant No. 101091374 of the MultiMiner project: Multi-source and Multi-scale Earth observation and Novel Machine Learning Methods for Mineral Exploration and Mine Site Monitoring.

Accuracy assessment is the final step in the development process which results in assessment for the quality of the application demonstrations done in the MultiMiner project. The Deliverable 4.2 &ldqu…

Accuracy assessment is the final step in the development process which results in assessment for the quality of the application demonstrations done in the MultiMiner project. The Deliverable 4.2 “Report on evaluation of the novel products in terms of their accuracy and applicability in mineral exploration and mine site monitoring at the respective MultiMiner test sites” is a contribution of MultiMiner Task 4.2:” Thematic Accuracy Assessment Methods“. The Deliverable presents the evaluation of accuracy and applicability of novel Earth Observation (EO) and machine learning-based products developed within the MultiMiner project for mineral exploration and mine site monitoring across five European test sites. This is a Deliverable of the MultiMiner project. The MultiMiner project is funded by the European Union’s Horizon Europe research and innovation actions programme under Grant Agreement No. 10109137474.

Conference abstract presented at the GeoInnovation Session at the 2025 OECD Conference of Mining Regions and Cities, 16-18 June, University of Lapland, Rovaniemi, Finland. GTK Open File Research Repor…

Conference abstract presented at the GeoInnovation Session at the 2025 OECD Conference of Mining Regions and Cities, 16-18 June, University of Lapland, Rovaniemi, Finland. GTK Open File Research Report 39/2025. 2025 OECD Conference of Mining Regions and Cities: The GeoInnovation Session – 16–18 June 2025, University of Lapland, Rovaniemi, Finland

The MultiMiner storymap presents a comprehensive overview of a pan-European initiative that harnesses multi-source and multi-scale Earth Observation to support responsible mining and environmental tra…

The MultiMiner storymap presents a comprehensive overview of a pan-European initiative that harnesses multi-source and multi-scale Earth Observation to support responsible mining and environmental transparency. Funded by the EU’s Horizon Europe programme, the project integrates in-situ measurements, UAV and airborne surveys, and satellite constellations to monitor the entire mining life cycle—from low-impact exploration to closure and rehabilitation. By combining advanced, non-invasive sensing techniques with innovative machine learning, MultiMiner reduces reliance on traditional ground-based drilling and point sampling while delivering continuous, spatially explicit information on mining-impacted landscapes. A central focus of the storymap is the development of novel spectral and geochemical monitoring methods for acid mine drainage (AMD) and mining-impacted waters. Using legacy and active mining sites in Kirki and the Kassandra district (Greece), the project builds a full-range (350–15,375 nm) spectral library of real AMD mixtures, couples UAV-based hyperspectral water surveys with detailed laboratory geochemistry, and employs Partial Least Squares Regression (PLSR) to quantitatively link reflectance spectra with concentrations of iron, sulfate, trace metals, and total suspended solids. These workflows enable safe, repeatable, and scalable mapping of contamination plumes and provide transferable protocols for extending UAV-derived relationships to satellite platforms. The storymap also documents a multi-tiered framework for air-dust monitoring that combines long-term PM10 records, passive dust samplers, vegetation spectroscopy, UAV hyperspectral imaging, and satellite products such as MODIS AOD, CAMS, EMIT, PRISMA, and ECOSTRESS. Case studies from Chalkidiki (Greece) and Ihalainen (Finland) illustrate how dust source areas, transport patterns, and deposition effects on vegetation can be jointly analyzed and how machine learning models (e.g. Prophet, XGBoost, Random Forest) can be used to forecast PM10 levels for operational dust surveillance. All key datasets, including open-access spectral libraries, water geochemistry, and validated modeling workflows, are designed to support reproducible research and to provide a blueprint for cost-effective environmental monitoring across mining regions in Europe and beyond. The presented StoryMap was conducted under the support of the EC through the MultiMiner project, funded under the European Union’s Horizon Europe research and innovation programme (Grant Agreement No. 10109137474).

The concentrations of arsenic (As) and sulphate (SO4) are important indicators in the assessment of surface water quality, particularly in regions with active mining operations. This study investigate…

The concentrations of arsenic (As) and sulphate (SO4) are important indicators in the assessment of surface water quality, particularly in regions with active mining operations. This study investigates the application of remote sensing techniques for the detection of As and SO4 in water bodies using Earth Observation data. We present a case study conducted on the Chalkidiki peninsula in Greece, encompassing three active surface mines—Olympias, Stratoni, and Skouries—operated by Hellas Gold, a subsidiary of Eldorado Gold Corporation. For this investigation, we employed a small and lightweight VNIR hyperspectral sensor (STS Spectrometer by Ocean Optics, with a spectral range of 337 – 823 nm) mounted on a DJI Phantom 3 drone to capture spectral data at 15 selected locations covering three streams related to the each of the studied mines. The hyperspectral data collection was complemented by in situ surface water sampling, followed by comprehensive laboratory analyses performed at Hellas Gold’s facilities. Due to the common limitations in direct identification of As and SO4 using imaging spectroscopy methods, the focus of our study was on indirect detection of the optically active water substances. For such purpose we utilized in situ measured parameters which showed strong correlation with the presence of these substances and could also be identified spectrally at the same time. Specifically, we targeted Total Suspended Solids (TSS) which strongly correlates with the stream flow. TSS exhibited statistically significant correlation with As concentration (in situ sampling data: R2 = 0.434) and at selected sites with SO4 concentrations (long term water quality measuring: R2 = 0.471 or higher). All these correlations were validated using the Hellas Gold’s regular water quality monitoring data, publicly accessible for the Chalkidiki peninsula (https://environmental.hellas-gold.com/). To estimate TSS (direct estimation) and As, SO4 (indirect estimation) using the acquired hyperspectral UAV spectral data we employed spectral indices (NDVI, NDWI1 and WRI1) adjusted for our spectral data characteristics (e.g., wavelengths of local spectral max/min). These unique spectral features were found between 580-590 nm (green-band), 670-680 nm (red-band) and 708-715 nm (NIR-band). The preliminary results indicate that the proposed spectral ratios, which are sensitive to spectral changes in water caused by TSS, could serve as proxies for estimating concentrations of As and SO4 in surface waters. This study demonstrates the potential of integrating drone-based hyperspectral imaging with in situ measurements for water-quality monitoring in mining-impacted regions, contributing to more effective environmental management practices. Acknowledgments: The presented analysis was conducted under the support of the EC grant MultiMiner. The MultiMiner project is funded by the European Union’s Horizon Europe research and innovation actions programme under Grant Agreement No. 10109137474.

Surface mining can have a significant impact on various aspects of the environment, including human health. One of the most harmful effects on the human body comes from particulate pollution, which is…

Surface mining can have a significant impact on various aspects of the environment, including human health. One of the most harmful effects on the human body comes from particulate pollution, which is dispersed into the air through mining processes. These pollutants, known as particulate matter (PM10), are particles smaller than 10 micrometers that can be inhaled and are associated with respiratory and cardiovascular illnesses. Monitoring this type of pollution is challenging due to its spatial distribution. The atmosphere is a three-dimensional space, and it is impossible to cover an entire locality with in-situ monitoring stations. However, remote sensing satellite data offers an effective alternative for monitoring airborne dust. Although satellite-based data may have limitations, they provide stable, spatial, and temporal coverage that is valuable for long-term monitoring and offers advantages in monitoring over large, inaccessible areas. This study has been carried out in Chalkidiki peninsula, Greece encompassing three active surface mines—Olympias, Stratoni, and Skouries—operated by Hellas Gold, a subsidiary of Eldorado Gold Corporation. Three Earth Observation (EO) datasets were used: the PM10 variable from the CAMS European air quality analysis dataset provided by the Copernicus Atmosphere Monitoring Service (CAMS), Aerosol Optical Depth (AOD) data at 0.55 micrometers from NASA’s MODIS Terra and Aqua satellites, and data from NASA’s SEVIRI sensor, where the Dust product was used for visual evaluation and the creation of a dust index. In-situ measurements (2015-2023) from Hellas-Gold dust monitoring stations (https://environmental.hellas-gold.com/) were used for building and validating spatial and temporal models. CAMS and MODIS data were aggregated into daily values for further evaluation. The CAMS data showed strong correlations with in-situ data, making them a valuable resource for monitoring, with correlation coefficients (R) around 0.8, depending on the locality. In contrast, MODIS data presented a different scenario, with correlations peaking at 0.5 and in some cases turning negative. This incompatibility may be attributed to long periods of missing data due to cloud coverage, so different MODIS preprocessing levels will be further examined. CAMS data was also aggregated into monthly and yearly values to observe trends, which indicated a decreasing tendency in concentrations, suggesting an improvement in air quality over the years. Additionally, machine learning methods, including Random Forest, LSTM, and XGBoost, were applied for data modeling and prediction. The XGBoost model produced the best results, with R values above 0,65 slightly varying within the sites of the interest. Future steps in this research will include further data modeling and prediction methods, with an emphasis on integrating artificial intelligence to develop an optimal workflow for this use case. Additionally, this methodology will be tested on another locality in Finland. ACKNOWLEDGEMENTS The presented analysis was conducted under the support of the EC grant MultiMiner. The MultiMiner project is funded by the European Union’s Horizon Europe research and innovation actions programme under Grant Agreement No. 10109137474.