Ground Penetrating Radar for the Exploration of Complex Mining Contexts

Abstract

Mining waste management is a significant environmental challenge that requires effective technical and economic solutions. In this context, the use of underground storage systems is sometimes a viable option to isolate this type of mining waste from the outside (especially when it poses a risk of environmental contamination). Despite the applicability and advantages of these structures, it is crucial to conduct thorough monitoring of the isolation and containment measures implemented during their construction. This study demonstrates how ground penetrating radar techniques can provide valuable insights into subsurface insulation layers with the aim of detecting potential water accumulation at depth and verifying the integrity of the seal and the state of buried materials. The results of the georadar survey applied on a mining case study demarcate the areas that should receive more attention in the near future and contribute to defining the most urgent actions to be implemented at the mining site. Beyond the evaluation of the 2D profiles, the research culminates in the creation of a 3D visualization tool for the entire mining site and its insulation layer, enabling users to inspect the structure’s condition at any location and obtain accurate depth measurements.

1. Introduction

Geophysical prospecting is highly valued for its time effectiveness and ability to characterize the different horizons of the ground without involving the alteration of its properties. These non-destructive techniques have become very popular in recent years due to their application and versatility in disciplines of diverse nature of the earth science [1,2]. Most of these methods were initially developed for ground deep exploration in order to obtain an approximate knowledge of the distribution and state of deeper geological structures. However, several of these geophysical techniques, especially the electromagnetic ones, allow their application at shallow depths, in fields such as the exploration of soils, roads or concrete structures [3,4]. In this sense, Ground Penetrating Radar (GPR) also known as surface penetrating radar is an active shallow geophysical technique that relays electromagnetic signals into the ground for subsurface investigation [5].

Among the numerous applications attributed to GPR methodology (described in the following section), the exploration of mining contexts and mineral deposits with complex geological conditions is one of the most noteworthy. Mining activity has been a long-standing key player in the supply of essential raw materials and society’s economic development. However, this industry has to commonly deal with elements of special interest, which can cause hazardous effects in the human population and ecosystem after prolonged periods of exposure to considerable levels [6,7]. For this reason, it is common to apply geophysical methods such as the mentioned GPR that contribute to improving the general subsurface characterization. The implementation of technology sometimes involves identifying mineral elements with potential for future extraction or exploitation, or evaluating the structural state of the subsoil, in terms of density, possible presence of cavities or water accumulation, among others [8,9,10]. In fact, enhancing the information content in mining exploration is mandatory to improve the efficiency of the operations involved, for which techniques such as GPR represent a significant contribution in expanding the understanding of the subsurface, providing valuable data for more informed decision-making.

Focusing on the aforementioned mining scenario, the implication of surface electromagnetics may in turn be related to the prospecting of scenarios that have already been restored and closed, but which require the periodic control of the remediation measures implemented at their origin. In this context of post-mining land management, long-term monitoring of restored sites is essential for ensuring environmental safety and regulatory compliance. Once the specific reconstruction and restoration measures of the ecosystem in a mining area (such as soil reconstruction, landform reshaping, re-vegetation or landscape restoration) have been applied, monitoring the status of the tailing structure is crucial, especially when mining activities were related to minerals of hazardous nature for the surrounding environment [11,12].

Expanding on the above, the present study focuses on the exploration of a mining deposit located in the Lavrion Technological and Cultural Park (Greece). The nature of the materials treated as part of the historical mining activity in the area, demands an exhaustive evaluation of the containment measures that were implemented to isolate the original mining waste. The purpose of this research is thus to explore the effectiveness of GPR on predicting the state of different structural layers used as a means of containment and isolation of deep mining waste buried after the abandonment of mining activities. The approach is based on the deployment of variable frequency GPR antennas (40–3440 MHz), selected to optimize the trade-off between penetration depth and spatial resolution. A series of parallel and orthogonal survey lines were acquired across the area of interest, enabling a three-dimensional reconstruction of subsurface features. The GPR data were processed using standard steps (including time-zero correction, background removal and bandpass filtering), followed by advanced migration algorithms to improve image fidelity and depth accuracy. Special attention was given to the dielectric properties of the subsurface materials, as variations in permittivity and conductivity significantly affect wave propagation and reflection characteristics. The study also integrates the GPR data with ancillary information to enhance interpretative reliability and facilitate the identification of buried interfaces, anomalies and potential degradation of engineered barriers. As can be seen below, the following subsection provides an overview of the primary applications of GPR technology, followed by a detailed description of the mining case study and the methodology employed, as outlined in Section 2. Section 3 highlights the most significant findings from the geophysical survey, which are then discussed in terms of their contribution to the characterization of the mining environment. Finally, the research concludes with a summary of the key findings and implications of the study. The insights obtained are expected to support future applications of GPR in the monitoring and management of legacy mining sites, contributing to the development of standardized diagnostic tools for evaluating buried containment structures. Moreover, the methodology presented here may serve as a reference framework for environmental risk assessment and remediation planning in similar post-mining contexts.

Review of Ground Penetrating Radar Applications

Ground Penetrating Radar constitutes a non-destructive, rapid and high-resolution geophysical technique with great potential for the characterization of the first horizons of the underground, providing 2D or 3D radar images of the subsurface [13]. This time-dependent electromagnetic technique has been developed over the past 40 years, mainly to investigate the shallow subsurface of the Earth in different applications and fields of study.

The operational basis of this technology can be found in the extensive published literature. In summary, it is based on radiation by a transmitting antenna of an electromagnetic wave that travels through the material at a certain velocity, depending on the electrical properties of the medium. If the wave passes through an object or a boundary layer between two materials with different electrical properties, part of the energy of that wave is reflected on the surface. This reflected wave is captured through the receiving antenna that is also part of the GPR equipment [14,15].

The ease of its use and the non-destructive nature of the technique make it widely applied for different purposes and in numerous fields of application. In this way, GPR is being used for military purposes such as detecting tunnels in demilitarized zones, identifying buried projectiles or accurately locating the remains of victims that were buried under a certain basement [16,17]. In the case of civil engineering applications, it has positioned itself as one of the preferred techniques, especially when it comes to detecting buried utilities as a prior step to more invasive work actions such as excavation. Other uses in this context include the inspection of transport infrastructure such as roads, highways or railways and the identification of possible deficiencies and damages in them, the existence of water or presence of specific metals, among other many possibilities [18,19,20]. Regarding archaeology and cultural heritage, Ground Penetrating Radar is also one of the most interesting methodologies for subsurface prospecting of archaeological sites or noteworthy heritage buildings. In a summarized way, its use in the field stands out in the inspection of the state of conservation of monumental buildings or the reconstruction and 2D or 3D visualization of buried archaeological ruins [21,22]. Other additional uses of GPR technology refer to the forensic and security fields, being applied in rescue operations, natural disasters or emergency intervention situations where the technique provides quick visual information of possible buried structures or vital signs or movements [23,24].

Beyond all the points mentioned above, one of the most promising applications of GPR lies in the geological and mining sectors. Considering the large contrasts of dielectric signals between the bedrock and the overburden, this methodology is positioned as an effective tool in the detection of the interface between layers, constituting a support to define the stratigraphic structure, the geometry of the rock body or the correlation among structures and sedimentary formations [25,26]. Several applications in this sense are found in the literature in paleo-seismological studies, with great results in the tracing of fault semi-nodes along strikes, also contributing to the determination of their chronology [27]. In the mining field, the detection of fractures of different size, from millimetric up to a decametric scale, is also a wide topic of GPR, being able to characterize rocks masses and mining sites in the sense of improving extraction and quarry processes and the definition of unstable rock slopes [28,29]. Whether the survey aims to reconstruct intricate subsurface geometries or identify variations in soil permittivity properties, the methodology plays a pivotal role in the field of near-surface geophysical techniques. Another notable use in this same sector is the capacity of the technique to determine openings and fracture fillings, essential in environmental applications and possible geothermal use, since these fractures sometimes represent the primary conduits for the transport of the hot fluid, providing valuable information on the primary origin of the geothermal resource [30,31].

Due to the versatility of ground penetrating radar and its effectiveness in geophysical prospecting, as previously mentioned, it has been used in this study for the investigation of the subsoil of the mining context described in the following Section 2. As stated before, while GPR has been extensively applied in mining contexts for subsurface imaging, structural characterization and void detection, most existing studies focus on either 2D profiling or isolated survey campaigns without long-term monitoring integration. In contrast, this study presents a novel approach by integrating GPR data with three-dimensional geospatial modeling and continuous monitoring protocols tailored to geomembrane and geotextile characterization within mining reservoir containment systems. This combined framework enhances the spatial-temporal resolution of subsurface diagnostics, offering a dynamic understanding of infiltration processes and material continuity. Unlike previous applications that often remain limited to qualitative assessments or local-scale interpretations, the present work advances toward a systematic, model-assisted interpretation of electromagnetic anomalies, thus contributing a replicable methodology for infrastructure integrity assessment under operational conditions.

2. Materials and Methods

2.1. Mining Context Under Study

As commented before, this work focuses on a mining site located in the coastal town of Lavrion, in the south-east of the Attica region of Greece. Geographically, it is located in a bay surrounded by mountains and characterized by its rugged coastline. This strategic emplacement was fundamental for the mining development of the area, its maritime accessibility being a great support for the transport of minerals. The coastline was also essential for establishing ports that supported commercial activities, particularly during the height of Athens’ prosperity in the 5th century BC. The mineral wealth of the region was recognized in ancient times, with mining activities beginning to develop in the 6th century. The area, renowned for its silver mines, played a key role in the resurgence and economic growth of Athens.

The importance of the mining development in the region is a direct consequence of the geological distribution that characterizes the area. Lavrion lies within the western part of the Attic-Cycladic metamorphic belt, situated in the back-arc area of the active Hellenic subduction zone. From the Eocene to the Miocene, the metamorphic rocks of the region (primarily marbles and schists) underwent multiple phases of metamorphism and deformation due to the collision and subsequent collapse of the Cycladic belt. During the Miocene, exhumation occurred through the activity of a large-scale detachment fault system, which also facilitated the intrusion of magmatic rocks and the formation of the renowned Lavrion silver deposits [32,33]. The area surrounding the mines features stacked nappes, with ore deposition concentrated in marbles, particularly at marble–schist contacts and in zones below, within or above the detachment fault. Remarkably, the Lavrion deposit comprises five distinct yet genetically related styles of mineralization, a phenomenon unique among known ore deposits, and hosts the widest variety of elements identified in any mining district worldwide.

The local geology, along with tectonic and magmatic activity and other external factors such as sea-level fluctuations over the last million years, significantly influenced surface oxidation processes. This oxidation led to the formation of an unparalleled variety of secondary minerals. As a result, the region contains 638 identified minerals, with Lavrion serving as the type locality for 23 of them (accounting for nearly 12% of all known mineral species). Renowned not only for its historical silver exploitation, the district also boasts a greater diversity of minerals than any other mining region on Earth [34].

The mineralogical diversity of the region encouraged strong mining extraction during the 19th and 20th centuries. Important mining companies were established in the area, being instrumental in the Lavrion development, driving the Greece industrialization and transforming the workers settlement into a town of 10,000 by the early 20th century.

The initial mining activities included mechanical processing facilities, hydro-mechanical ore beneficiation units and furnaces for thermal ore processing and lead reduction. Ores extracted from the mines were first beneficiated on-site. Lead, zinc and mixed sulfide ores underwent mechanical preparation using dollies and ore-washing units before metallurgical processing. In 1905, large-scale modernization of lead metallurgy began, introducing two new calcination methods tailored to ore types, and expanding the metallurgical activities to include further processing of the reducing smelting products. After a period of growth and technological development, during the 1970s and 1980s, a crisis struck Greece’s key industrial centres, including Lavrion, where after over a century of operation, its mining activities ended [35].

In an attempt to preserve the old mining facilities of the region, the concept of creating a technology park and a museum of mining and metallurgy to safeguard the site’s historical and technological significance is introduced. In this sense, the National Technical University of Athens (NTUA) created the Lavrion Technological and Cultural Park (LTCP), established since 1995 with funding from both EU and national sources.

The industrialization of Lavrion led to significant environmental degradation due to the intensive mining and metallurgical activities. As a result, some of the key objectives of LTCP include aiding in the decontamination of the area and the safe treatment of mining waste generated over the years. In this line, one of the projects of the park involved the excavation, transportation and safe disposal of polluted soils from different locations within the LTCP boundaries. These soils were deposited at a specially prepared hazardous waste landfill site known as “Dry Tomb”, designed to meet the highest technical and environmental standards, and to contain mining waste that poses significant risks to human health and the surrounding ecosystem (Figure 1). Figure 1 shows in (A) the general view of the region of Lavrion within the country of Greece and in (B) the perimeter of the mining deposit included as case of study.

As shown in Figure 2, the mining and metallurgical waste was deposited in a stockpile isolated from the outside environment by a series of protective layers. Once the construction process of the site was completed, it was permanently closed, sealed and remains under continuous environmental monitoring to ensure its safe and compliant operation.

A total waste volume of 130,000 m³ was deposited in the stockpile, being the waste excavated and transported from multiple locations within the LTCP boundaries. The typical chemical composition of the waste in the repository, based on the analysis conducted by the park, is presented in

Table 1. Average metal and metalloid concentrations in the isolated stockpile (mg kg⁻¹).

Element	Average Concentration	Maximum Concentration
Pb	32,430	67,700
As	4517	10,800
Cd	160	600
Cr	4	10
Ni	150	300
Zn	47,843	97,900
Mn	9002	16,600
Fe	154,800	275,200
Cu	2548	6300

.

2.2. Work Methodology

The methodological process followed in this research includes a prospective campaign performed with the Ground Penetration Radar Proceq GS8000 commercially supplied by Screening Eagle (Schwerzenbach, Switzerland). The objective of this geophysical survey is to obtain an approximation of the state of the containment and isolation structures used in the considered mining deposit. Detailed information on the GPR equipment is provided in the following

Table 2. Technical specifications of the GPR used for the prospecting campaign performed in this research.

GS8000 Technical Specifications
Radar technology	Reduced Frequency Continuous Wave
Modulated frequency range	40–3440 MHz
Effective bandwidth	3200 MHz
Minimum detectable target size	1 cm
Maximum depth of penetration	10 m
Scan speed	500 Hz
Spatial interval	Up to 100 scans/m
Acquisition speed	Up to 80 km/h
GNSS receiver	Multiband GPS + Glonass + Galileo + Beidou

.

Data collection in the mining stockpile was carried out in adaptation to the surface conditions of the deposit, where the existence of vegetation of a variable height did not allow for regular meshing. In this approach, multiple free passes were made (depicted in Figure 3) covering the maximum possible area of the deposit to ensure comprehensive data collection and subsurface analysis. GPR equipment is multi-frequency, operating within a range from 40 to 3440 MHz, that provides a different resolution and penetration capacity depending on the applied frequency. Frequency is adapted according to the need to balance penetration depth and spatial resolution across heterogeneous subsurface conditions. Lower frequencies allow for deeper penetration, essential for identifying the basal structure and potential deep water infiltration paths, while higher frequencies enhance the resolution for detecting thin-layer features. This broadband approach aligns with established practices in GPR subsurface mapping over complex lithologies and enables multiscale data fusion.

The survey layout employed a semi-structured free-line acquisition strategy to adapt to site constraints and ensure maximal surface coverage. While this method introduces challenges in interpolation due to non-uniform line spacing, high-density point acquisition combined with geostatistical kriging and variogram modeling allows reliable volumetric reconstruction, as validated in prior studies using similar adaptive schemes in mining and environmental contexts [36].

Environmental interference, particularly from surface vegetation and moisture variability, was mitigated through post-processing filters to correct for signal attenuation and scattering. Signal correction techniques included gain adjustments and background removal based on standard radargram analysis procedures, along with dielectric property calibration using reference targets. These methods ensured consistent signal interpretation even under variable surface conditions.

Data Processing

After the field campaign, the data were subsequently processed in the web-based software GPR Insights of Screening Eagle. The tool is conceived for complex dataset analysis and visualization, delivering superior GPR data post-processing quality owing to the long-established algorithms and advanced processing methods [37]. In the case of the present work, GPR data were processed in 2D (radargrams) by applying different filters such as zero time, automatic gain, Hilbert transform and background noise removal, in order to obtain a clean signal and facilitate the subsequent results interpretation.

In addition, 3D visualizations were generated with the same tool to provide an overview of the deposit and the current status of the containment measures. However, given the limitations of the collected registered data, especially in central areas, and in order to have a continuous data model, an interactive visualization platform based on different interpolation processes from the available GPR records was developed using ArcGIS Pro 10.

For spatial interpolation, ordinary kriging was applied using 90% of the available data for training and reserving the remaining 10% for validation. Prior to interpolation, an empirical multiplicative skewing transformation was applied to correct for data asymmetry and approximate a normal distribution, which is a typical assumption for many geostatistical methods. The experimental semivariogram was fitted using a spherical model, which reflects strong spatial correlation at short distances that gradually diminishes with increasing separation. The interpolation model was evaluated through cross-validation, yielding a root mean square error (RMSE) of 0.17008 (a priori). Subsequently, the “GA Layer To Points” tool was used to compare predicted values against the reserved validation dataset. The squared differences between predicted and observed values summed to 3.68097 over 117 points, resulting in an a posteriori RMSE of 0.17737. The small difference between the a priori and a posteriori RMSE confirms the accuracy and reliability of the interpolation model (Figure 4) [37,38,39,40,41].

The interpolation results show that the mean square error is 0.17009. To ensure that these results are correct, the “GA Layer To Points” option of ArcGIS Pro is used on the interpolation, using the points set aside for validation (10% validation testing). This process provides the difference between the actual and interpolated values. The final mean square error obtained was 0.17737, thus validating the interpolation process implemented on the data model.

The overall workflow of processing the ground-penetrating radar data has been summarized in the following Figure 5.

3. Results

The most significant results of the GPR survey carried out in the mining case under study are presented below. Initially, the 2D radargrams corresponding to the most significant records for the study’s objectives are presented. These profiles were selected based on their relevance to the subsurface features of interest. Subsequently, a 3D visualization of the entire deposit is generated by integrating all GPR profiles, allowing for a detailed and spatially coherent representation of the subsurface structure. This approach facilitates a more comprehensive interpretation of the deposit’s geological and mining-related characteristics.

3.1. 2D Profiles

The 2D profiles obtained from the processing of the GPR data in the aforementioned tool show a fairly adequate resolution and highly valuable information. To improve the scale and facilitate the visualization of the 2D profiles, each of the field registers (Figure 3) has been divided into sections of approximately 20 m. As can be seen, greater emphasis has been placed on those profiles where the position of the geotextile is variable and phenomena of water accumulation and infiltration are found. In addition, the evolution of the first layer of geotextile (probably the only one visible in the radargram) has been marked in each profile, as well as the possible complex areas due to the aforementioned accumulation of water.

• Registers 1 and 2

When these profiles are analyzed, a marked reflection is noted (in green in Figure 6), which indicates the change in the electromagnetic properties of the layer with respect to the surrounding material. This suggests that these two profiles are characterized by a slight intuition of the first geotextile layer, at a fairly constant depth of around 0.9 m. On the other hand, no significant changes or reflections are found in the rest of the profiles, denoting that no major distinctive phenomena that could be useful in the characterization of the deposit are observed, so only an example of profile 1 is shown in the first section (Figure 6).

• Register 3

As in the previous cases, the profile contains a marked reflection that would correspond to the first layer of the geotextile material (with an irregular distribution along the profile), with different electromagnetic properties than the materials of the surrounding sub-use material. Areas with concentrated reflections are also observed in different parts (marked in the blue square), which, given their position and form of visualization, suggest the possibility of water infiltration phenomena (Figure 7, Figure 8, Figure 9, Figure 10 and Figure 11).

• Register 4

This register shows a clear marking of the geotextile layer (shown in green), appearing quite attenuated and at an approximately constant depth until the middle of it, where the layer rises to surface levels. Additionally, the irregularity of the reflections at certain points in depth gives an idea of the possible concentration of water derived from possible infiltration processes (Figure 12 and Figure 13).

• Registers 5 and 6

The electromagnetic response obtained from these profiles is not fully conclusive, as they were recorded in the central area of the deposit, where the presence of taller vegetation interferes with the proper transmission and reception of the GPR signal. Figure 14 includes one of the profiles where the geotextile layer can be seen in some areas with fairly stable behavior and at a depth of around 1 m. Areas with probable water accumulation phenomena are also detected, given the presence of different electromagnetic responses in various areas of the profile (blue squares).

• Register 7

As in the case of register 3, and in line with its location in the deposit, this radargram shows a marked evolution of the first layer of geotextile with interesting results that denote its variability and its proximity to the surface at some points. The most significant 2D profiles are shown in the following Figure 15 and Figure 16.

• Register 8

The area covered by this profile shows apparent continuity of the geotextile at a depth of approximately 1 m (as in the initial construction). The shape of the electromagnetic response and its different concentration in certain parts of the profile reveal areas with possible infiltration and accumulation of water with certain attenuation of the signal (Figure 17).

• Register 9

This register covers one of the external areas of the deposit with interesting results that demonstrate the irregularity of the distribution of the first layer of the geotextile, with a marked appearance almost on the surface in some points. As in the previous cases, the different electromagnetic nature of the geotextile is translated in the radargram as a marked reflection that allows its differentiation from the rest of the materials. As presented in the following Figure 18 and Figure 19, some water infiltrations are also found.

• Register 10 and 11

The two profiles crossing the central area of the reservoir reveal a nearly continuous geotextile layer located at the depth initially projected, as evidenced by a consistent variation in the electromagnetic response attributable to the contrast in dielectric properties between the geotextile and surrounding materials. Moreover, localized anomalies indicating potential water accumulation are observed in specific sections, likely associated with differences in moisture content and dielectric permittivity, which produce distinct signal attenuations and phase shifts (Figure 20).

• Register 12

An area characterized by distinct zones of predominant water infiltration, with no significant alterations detected in the spatial continuity or distribution of the geotextile layer, as inferred from the uniformity of the electromagnetic response (Figure 21).

3.2. 3D Modelling

Alternatively, a 3D model has been generated using the GPR equipment software (Insights), by integrating the complete set of radargrams obtained from the study area. This model serves as a valuable tool for enhancing the evaluation of the results, allowing for a more comprehensive interpretation of the subsurface conditions.

By providing a continuous representation of the data, the 3D model (presented in Figure 22) helps to visualize spatial relationships between the detected anomalies. This approach improves the overall understanding of the deposit’s structure and facilitates the identification of patterns that may not be as apparent in individual radargrams.

Furthermore, the model enables a more detailed assessment of problematic areas, supporting decision-making processes related to isolation measures, drainage improvements and future maintenance strategies. The integration of 3D visualization into the analysis thus contributes to a more effective and data-driven approach to managing the stability and integrity of the deposit.

In addition to the above, the results of the data interpolation have allowed the geotextile layer to be modeled in 3D and included in a three-dimensional visualization platform. For this purpose, rendering was performed using the open-source library Potree^® (http://potree.org/, accessed on 20 December 2024). This platform is built on WebGL technology and can efficiently render large point clouds while supporting metric operations such as linear measurements, area and volume calculations, which may be valuable to professionals. The integration between this engine and the virtual tour was achieved through an HTML page embedding the 3D point cloud. Additionally, this viewer (Figure 23) includes various JavaScript functions that enable retrieving point coordinates, linear measurements and area values for specific locations [42]. The 3D viewer can be consulted using the following link: https://iedu.usal.es/clouds/lavrio/ (accessed on 10 January 2025). This tool enables users to visualize the entire geotextile layer and the site, create profiles and analyze the distribution of the insulation layer. Additionally, this information can be exported in CAD format. Given the context of a mining waste storage facility, direct ground-truth data acquisition has not been possible due to safety, operational and environmental constraints. Consequently, the validation of the 3D visualization and interpretation tool has been limited to internal consistency checks, such as interpolation error metrics and coherence with expected stratigraphic patterns. However, the results of this study are intended to serve as a basis for future field campaigns, where targeted sampling and in situ measurements will be planned based on the geophysical anomalies and structural hypotheses derived from the current GPR dataset.

4. Discussion

4.1. Status of Containment and Isolation Layers

The evaluation with ground-penetrating radar in the case study presented here has provided an important contribution to understanding the status of the containment measures that were originally implemented at the mining site. Despite the limitations of this electromagnetic technique in the exploration of medium and deep underground levels, it can still provide valuable information in applications such as the one analyzed here. In such cases, it offers a relatively simple and rapid method for acquiring essential data, contributing significantly to the overall subsurface characterization.

In this sense, GPR has allowed an evaluation of the conditions of the first layer of geotextile used in the construction of the mine waste insulation system. As shown in Figure 2, the structure was initially made up of a series of layers and drainage materials to ensure the complete isolation of the deposit from the outside. As can be deduced from the radargrams presented in the Section 3, the reflections and alterations observed only provide information on the location of the first layer of the geotextile. Despite this, the new knowledge provides added value in defining future actions and maintenance that should be provided to the deposit, especially in those parts where this layer is found practically at superficial levels.

In this way, and based on the results obtained with the GPR, it has been possible to identify the most problematic areas of the deposit that should be taken into special consideration in its future treatment. Figure 24 shows the locations where the first geotextile layer has been located at a nearly superficial level (marked in red) and those where this layer appears at subsurface levels but far from its original position (in orange).

As observed, the geotextile layer appears to be in a state of degradation, particularly along the eastern perimeter of the site. This deterioration, likely due to prolonged exposure to surface conditions, may be compromising its intended insulation function.

Conversely, the central area of the deposit, as well as the northwestern region, maintains a consistent and uniform distribution of the first geotextile layer. No significant displacements or alterations from its original placement have been observed in these areas, suggesting that the insulation function remains intact and has not been affected by external factors. This stability contrasts with the degradation observed in other sections, indicating that certain areas of the deposit have better preserved the integrity of the geotextile layer over time.

4.2. Possible Water Filtration Phenomena

Beyond identifying the location of the isolation measures, another key aspect of this investigation has been the detection of areas where water accumulation or infiltration may be occurring. The presence of moisture within the deposit can significantly impact its stability and the effectiveness of the geotextile layer, potentially accelerating material degradation and compromising its intended function.

The analysis of GPR results has made it possible to identify specific zones where anomalous reflections suggest the presence of water. These areas may be indicative of drainage deficiencies or breaches in the isolation system that allow water to penetrate and pool within the deposit. Such conditions can lead to structural weakening, increased erosion and potential contamination risks.

Understanding the distribution and extent of water infiltration is essential for developing appropriate mitigation strategies. Future efforts should focus on reinforcing drainage systems and implementing targeted maintenance measures to ensure the long-term effectiveness of the isolation barriers and the overall stability of the deposit.

In the same Figure 24, specific points have been marked where the evaluation of electromagnetic reflections indicates the potential presence of water at various depth levels. As observed, the majority of these water accumulations are concentrated along the perimeter of the deposit, while no similar phenomena have been detected in its central area. This distribution suggests that water infiltration is more likely to occur near the boundaries of the deposit, possibly due to external sources of moisture, surface runoff or weaknesses in the isolation measures at the edges.

4.3. Research Contribution and Correlation with Previous Geophysical Campaigns

One of the most relevant studies performed in the area under study is the evaluation of the geological composition and material distribution of the Lavrion deposit through an aerial magnetometry survey [43]. The mentioned research provided precise mapping of the site based on a high-resolution and non-invasive approach to capturing magnetic field data. The final result of this work provides a 3D magnetic susceptibility model that represents a detailed view of the magnetic susceptibility variations within the repository. This model allows the visualization of high-susceptibility areas associated with ferromagnetic materials and low-susceptibility zones linked to diamagnetic materials such as lead, arsenic, cadmium and zinc.

Although GPR and magnetometry are prospective methodologies of completely different natures and application purposes, it is interesting to consider the available information and the possibilities of complementarity. The results of ground-penetrating radar surveys, primarily for structural purposes, have identified the most affected sectors of the reservoir and those with the greatest degree of disruption to the isolation measures, as well as the location of potential water infiltrations. Regarding airborne magnetometry, it has provided an estimated distribution in depths of diamagnetic and ferromagnetic materials, which is especially useful in evaluating possible future ore extractions.

However, knowledge about distribution in the depth of these materials is also valuable for planning actions on the geotextile layer and assessing the need for its repair, based on the layer’s condition (GPR results), and the presence of complex or particularly contaminating materials at depth (magnetometry). Furthermore, the acquisition of both geophysical characterizations can significantly optimize the restoration of isolation measures and the potential extraction of target minerals. By integrating these datasets, it is possible to prioritize actions based on a detailed understanding of subsurface conditions, enabling a more efficient allocation of resources and aligning the restoration and extraction processes with the specific technical requirements of the mining site operation.

Thus, according to magnetometric prospecting, there is a greater concentration of materials with high magnetic susceptibility in the central and southeastern areas of the deposit, while the northern area has the lowest magnetic susceptibility values. By identifying the areas with the greatest wear of the geotextile layer and potential water infiltration, the planning of actions on the deposit can be approached jointly and based on the mineral considered of greatest interest (according to the current market), as well as any potential sources of contamination due to the nature of the material expected at depth.

Regarding the products derived from this research, it is clear that the precise characterization of the different components of the site insulation measures provides substantial insight beyond the expectations based on the original construction of the structure. Additionally, the 3D visualization tool serves as a valuable resource for accurately mapping the in-depth location of the geotextile material, also facilitating the evaluation of future interventions and the comparative monitoring of its condition over time.

In order to explore the relationship between the mentioned geophysical methods, a data fusion analysis is provided in the following

Table 3. Data fusion analysis between GPR and magnetic results.

Area	GPR Anomalies	Magnetic Susceptibility	Interpretation
Central Sector	Minimal GPR anomalies	Moderate to high	Intact geotextile layer, lack of correlation with ferromagnetic minerals
Southeastern Sector	High attenuation, disrupted geotextile	High	Correlation suggests possible chemical interaction or mechanical stress from ferromagnetic accumulation
Northern Sector	Moderate anomalies and infiltration signatures	Moderate to low	Localized structural degradation not necessarily associated with ferromagnetic mineral accumulations
Peripheral Zones	Localized discontinuities	Mixed values	Further investigation is required, possibly related to geomorphological variability

.

5. Conclusions

Mining waste represents a significant environmental issue that must be addressed from a technically and economically effective perspective. In this sense, underground deposits for mining waste storage are a viable solution for managing such waste, especially when it is hazardous or challenging to handle on the surface. In this context, it is essential to carry out exhaustive monitoring and control of the isolation and containment measures employed in the original construction of these structures. Therefore, this research demonstrates how a prospective study with ground-penetrating radar can provide valuable information for characterizing subsurface insulation layers and identifying the potential accumulation of water at depth. This information is crucial for verifying the integrity of the deposit’s seal and assessing whether the buried materials are adequately isolated from the atmosphere and potential surrounding bodies of water. Findings have indicated the most problematic parts of the mine in terms of the condition of the geotextile layer used as part of the insulation structure, as well as those areas most susceptible to deep water accumulation. The research culminates in the development of a 3D viewer of the entire mining site and the insulation layer, which offers the possibility of checking the condition of this structure at any point and taking precise measurements of its location at depth.

These conclusions reinforce the role of GPR as a simple and versatile geophysical prospecting technique from an operational point of view, which provides a basis of undoubted value in the characterization of subsurface levels, without compromising the integrity of the analyzed structure. This non-invasive method allows for the precise assessment of materials, their distribution and potential issues such as the ones addressed here, making it an essential tool in both environmental monitoring and resource management.

Despite the advantages demonstrated, several limitations of GPR technology must be acknowledged. GPR performance is highly dependent on subsurface material properties; signal attenuation in high-conductivity media such as clays or wet silts can significantly reduce penetration depth and resolution. In addition, depth limitations (generally up to several meters in optimal conditions) may hinder full characterization of deep structures in certain contexts. Heterogeneous layering and metallic interferences can also compromise signal clarity and data interpretation. In the specific case of this investigation, the signal is attenuated once it passes through the first layer of the geotextile and in areas where water accumulates, which prevents the deeper containment layers from being seen. To overcome these limitations, future work could integrate multi-frequency antenna arrays, which would allow for the combined benefits of shallow high-resolution imaging and deeper penetration. Likewise, the incorporation of UAV-mounted GPR systems may enhance accessibility to difficult terrain, improve survey efficiency and allow for large-area coverage with minimal disruption

In conclusion, while GPR proves to be a non-invasive, operationally simple and versatile tool for subsurface characterization in mining environments, its maximum effectiveness depends on the proper understanding of its physical limitations and the adoption of integrated strategies to maximize data reliability and spatial coverage. Additionally, its ability to provide real-time, high-resolution data supports more effective decision-making and long-term planning, ensuring that critical issues can be identified and managed promptly.