Conference Agenda

Arsenic (As) and sulfate (SO₄) concentrations are important indicators of surface water quality, particularly in areas influenced by mining activities. This study investigates the use of remote sensing to detect As and SO₄ in water bodies through Earth Observation (EO) data, focusing on a case study from the Chalkidiki peninsula in Greece, where are three active surface mines—Olympias, Stratoni, and Skouries (operated by Hellas Gold).

We employed a VNIR hyperspectral sensor (STS Spectrometer, Ocean Optics with a spectral range of 337–823 nm) mounted on a DJI Phantom 3 drone to capture spectral data at 15 sites across three streams linked to the mines. This data was complemented by in situ water sampling and laboratory analysis at Hellas Gold's facilities.

Due to the limitations in directly detecting As and SO₄ using imaging spectroscopy, we focused on indirect detection via optically active substances, specifically targeting Total Suspended Solids (TSS). TSS showed a substantial correlation with As (in situ sampling data: R² = 0.434) and, at selected sites, with SO₄ (long-term water quality measurements: R² = 0.471 or higher). These correlations were further validated using Hellas Gold's public water quality monitoring data (https://environmental.hellas-gold.com/).

To directly estimate TSS and indirectly determine As and SO₄ levels, a spectral sensitivity analysis was conducted using drone-based hyperspectral data. The red-band wavelength range of 670-680 nm exhibited the strongest correlation with TSS (R² = 0.512), leading to the calculation of spectral indices such as NDVI (R² = 0.442), NDWI1 (R² = 0.353), and WRI1 (R² = 0.403). In order to scale up the results from the drone-based data, equivalent indices were derived from satellite multispectral PlanetScope imaging data utilizing bands B4, B6, and B7. The PlanetScope data also displayed substantial correlations with TSS (for NDVI is R² = 0.561, for NDWI1 is R² = 0.659, and for WRI is R² = 0.668), highlighting its potential for broader spatial and temporal monitoring purposes.

Our results indicate that these indices, which respond to spectral changes in water caused by TSS, could serve as proxies for As and SO₄ concentrations. This study demonstrates the potential of integrating in situ, drone-based and satellite data for comprehensive water-quality monitoring contributing to more effective environmental management practices.

Spectral absorption feature parameters (SAFPs) derived from airborne hyperspectral imaging (HSI) (396.0–2453.5 nm) were used to predict Pb contents in coalfield overbank sediments. The derived SAFPs were associated with goethite (~500 nm) and kaolinite (~1448 nm) in sediments. Sediment Pb contents correlated strongest with goethite-related absorption-depth (r = 0.6). The calibration model had a R² = 0.69 and standard error of estimation (SEE) = 3.97, after outlier removal. The validation model had a R² = 0.65 and SEE = 3.90. Overall, the results suggest that airborne HSI can complement conventional geochemical methods of detecting Pb contents in overbank sediments.

According to official figures, aggregate extraction has intensified along the Guaviare River to meet Colombia’s growing infrastructure demands, raising concerns about its geomorphological stability and the socio-environmental wellbeing of riverine communities. This study combines multitemporal Landsat imagery (1984–2023) with official mining data to evaluate channel migration, erosion, and deposition across a 230 km river reach. Seven meanders were classified according to their distance from principal mining hotspots—high, medium, and low influence—revealing that meanders nearest to extraction sites exhibit up to 60–70 % higher migration rates and nearly double the erosion observed in more distant meanders. A distinct peak in deposition at a medium-influence meander (24.5 km downstream) further underscores the heterogeneity of fluvial responses, which are affected by both direct mining impacts and localized sediment accumulation. Temporal analyses demonstrate a threefold increase in erosion between 2008 and 2013, coinciding with the onset of heightened production (~2012), followed by elevated deposition from 2013 to 2018 and a renewed surge in erosion after 2018. Minimal correlation with deforestation or river discharge levels suggests that aggregate mining is the primary driver of these channel adjustments. However, potential unreported or illicit extraction beyond officially documented sites complicates the assessment, indicating that official records may underestimate the full extent of mining-related impacts. The results highlight an urgent need for integrated management and enforcement strategies that balance economic imperatives with ecological and cultural preservation along the Guaviare River.

Keywords: Alluvial mining, Guaviare river, Remote sensing, Gravel extraction, fluvial dynamics, river water

A significant proportion of the environmental challenges associated with mining are related to the waste produced during the extraction process, with pit lakes being one of the most difficult mining legacies to manage. These water bodies are susceptible to physico-chemical changes due to prolonged exposure to minerals (such as pyrite), which can create acidic conditions and increase sulphate and trace element concentrations in the pit lake and the surrounding water bodies. This study aims to develop a robust methodology and numerical dataset of these water bodies to analyse, monitor and determine the physico-chemical characteristics of pit lakes in Chile using satellite imagery.

The novelty of this research lies in its development of a low-cost methodology for monitoring pit lakes through satellite imagery and water colour analysis in the HSV (Hue, Saturation, Value) space. The true colour of water, influenced by the absorption, reflection, and transmission of specific wavelengths of sunlight through dissolved ions and suspended particles, serves as a proxy for understanding its chemistry. Transition metals create distinctive coloured complexes—such as olive-green for Fe²⁺ and blue-green for Cu²⁺—while fine colloids scatter sunlight in specific ways, revealing information about the water's pH and redox conditions. This study calculates time series of HSV values for each pit lake using surface reflectance data from Landsat and Sentinel-2, processed via the Google Earth Engine with Python. A supervised machine learning technique, Random Forest, was employed for the identification and segmentation of pit lakes over time, successfully leveraging the Normalized Difference Water Index (NDWI) to enhance the accuracy of the monitoring approach. Overall, this methodology provides a scalable solution for assessing the dynamic physico-chemical characteristics of pit lakes.In situ physicochemical parameters (pH, EC, Cu and Fe) were collected from publicly available documents. These two datasets are then integrated and compared.

The results of this study confirm that it is possible to obtain HSV parameters from pit lakes. Furthermore, the HSV parameters and their variations over time are directly associated with the physico-chemical characteristics of the pit lakes.

This method provides an accessible and efficient tool for continuous monitoring of water quality in pit lakes, without relying solely on costly and safety compromising in situ sampling campaigns. It has direct applications in the management of mines undergoing closure or already closed, enabling compliance with international environmental regulations. This methodology can be adopted globally, promoting sustainability in mining and protecting water resources.

In the past decade, unmanned aerial vehicles (UAVs), or drones, have become invaluable tools in the mining industry, supporting applications ranging from mineral exploration to environmental remediation. When equipped with high-resolution image sensors, UAVs enable high spatial and temporal resolution surveying, making them particularly effective for monitoring abandoned mining sites. This paper presents a case study demonstrating the application of UAV technology for monitoring an abandoned mining site affected by acid mine drainage (the Trimpancho mining complex in the Iberian Pyrite Belt, southwestern Spain).

The study was conducted on two selected waste dumps: one at the beginning and the other at the end of the mining complex. The primary objective was to estimate the volume of waste material accumulated in these areas. To achieve this, a DJI Phantom 4 RTK UAV was employed to conduct comprehensive aerial surveys of the sites. The imagery collected enabled the generation of orthophotomaps and digital surface models (DSMs), which provided detailed spatial information for accurately delimiting waste accumulations and identifying dominant runoff zones contributing to the degradation of the Trimpancho stream.

The three-dimensional models and orthomosaics produced from the UAV data provided an in-depth visualization of the waste distribution. This enabled precise volume estimations that are essential for long-term monitoring and management of the mining system. By comparing these models with future surveys, it is possible to track changes in waste dump morphology over time. Additionally, the models offer a valuable tool for assessing the potential valorization of critical materials accumulated in the abandoned dumps.

The results confirm that UAV technology is highly effective in obtaining the detailed and accurate data needed for monitoring abandoned mining sites. This information is crucial for planning and implementing remediation strategies adapted to the site's specific topography, hydrology, and geology. UAVs present an innovative, efficient solution that optimizes safety, accuracy, and cost-effectiveness. By enhancing assessment precision and enabling targeted reclamation strategies, UAVs technology contributes to the restoration and sustainable management of degraded mining areas, while also advancing environmental monitoring efforts.