Estimating Mercury and Arsenic Release from the La Soterraña Abandoned Mine Waste Dump (Asturias, Spain): Source-Term Reconstruction Using High-Accuracy UAV Surveys and Historical Topographic Data

Abstract

The waste dump from the abandoned La Soterraña mine, a former mercury extraction site, contains high concentrations of mercury (Hg) and arsenic (As), which pose a significant environmental risk due to direct exposure to the environment. Given the site’s topography and slope, surface runoff has been identified as the primary mechanism for the dispersal of these toxic elements into nearby watercourses. This study quantifies the amount of Hg and As released into fluvial systems through surface runoff from the waste dump. Historical topographic data, Airborne Laser Exploration Survey public data from the National Plan for Aerial Orthophotographs (1st PNOA-LiDAR) of the Spanish Ministry of Transport, Mobility and Urban Agenda, and high-precision photogrammetric drone surveys were utilized, with centimeter-level accuracy achieved using airborne GNSS RTK positioning systems on the drone. The methodology yields reliable results when comparing surfaces generated from topographic data collected with consistent methodologies and standards. Analysis indicates an environmental release exceeding 1000 kg of mercury (Hg) and 12,000 kg of arsenic (As) between 2019 and 2023, based on high spatial resolution data (GSD = 8 cm). These findings highlight a sustained temporal contribution of chemical contaminants, which imposes serious environmental and biological health risks due to persistent exposure to toxic elements.

1. Introduction

Abandoned mining sites represent persistent environmental and public-health risks due to the potential release of contaminants from waste deposits. Among these, former mercury mines are particularly concerning, given the toxicity and persistence of mercury in ecosystems [1]. Surface runoff from waste rock piles has been identified as a key mechanism for the mobilization and spread of these contaminants, transporting particles through erosion processes into nearby soils and water systems. Research addressing this process seeks to quantify the release rates and mechanisms of material transport from these sites, with the goal of informing effective monitoring and remediation strategies that can mitigate environmental impacts.

The present study examines material release from a mercury mine waste landfill, with a focus on quantifying the effects of surface runoff on erosion and contaminant dispersal. To achieve this, digital terrain models (DTMs) derived from classical topographic methods and drone-based photogrammetry are utilized to provide high-resolution, spatially accurate representations of surface alterations. Traditional topographic surveys offer well-established ground control, whereas drone photogrammetry allows for extensive spatial coverage and enhanced resolution with increased efficiency in time and cost. This dual-method approach allows for a more comprehensive analysis of the terrain changes associated with material transport, combining the precision of classical surveying with the flexibility of modern remote sensing techniques.

By integrating these methods, this research aims to address gaps in current understanding of erosion and material mobilization at legacy mining sites. The study further seeks to evaluate the efficacy of these techniques in capturing detailed topographic data and monitoring erosion processes over time. Findings are expected to contribute to a broader understanding of the environmental impacts of abandoned mining sites, providing insight into the use of digital terrain models for long-term monitoring and aiding in the development of targeted mitigation measures.

2. Materials and Methods

2.1. Study Area

The study area is located in northwestern Spain, Principality of Asturias, in the municipality of Lena, whose study case is the main waste dump of La Soterraña Hg abandoned mine, in the surroundings (Figure 1). It is located on the side of the local road AS-231 (Peñamiel—Pola de Lena), on the southwest face of Mount Campusas, located 600 m (horizontal distance) from the nearest residential building of Muñón Cimero. The Muñón stream, which flows into the Lena River, originates just a few meters below the mining site.

The studied site is an ancient mercury mine where cinnabar (HgS) and native mercury deposits were primarily extracted and processed, alongside materials from nearby extraction zones. The mineral paragenesis is mainly composed of cinnabar and native mercury, with lesser occurrences of realgar and pararealgar (As₄S₄), orpiment (As₂S₃), pyrite, marcasite, arsenopyrite, arsenical pyrite, and associated iron oxides [2]. The exploitation of this site dates back to Roman times, initially for gold mining purposes. [3]. During the 19th and 20th centuries, the mine alternated between periods of activity and inactivity, reaching its peak development between the 1940s and 1970s [4]. Mining operations at this site have resulted in waste materials with elevated concentrations of Hg and As.

Historical Evolution of the Site

The first mining registration application in Asturias aimed at exploiting a mercury deposit, registered as “La Peña” and located in the municipality of Mieres, dates back to 1838 (

Table 1. Mining Concessions Requested [4].

Year	Name	Nº	Location	Registrant
1838	Peña	–	La Peña (Mieres)	Manuel García Argüelles
1842	Argayos	18	Argayos (Mieres)	Antonio Cavanilles
1842	La Flecha	22	Valle Miñera (Mieres)	El Porvenir
1843	La Peña	23	La Peña (Mieres)	José María Cavanilles
1843	La Clave	49	La Peña (Mieres)	Luis Vigil Quirós
1843	Carba	246	Valle Miñera (Mieres)	La Virtud
1843	Ortigales	254	Valle Miñera (Mieres)	Soc. La Constancia
1843	Asturiana	288	La Peña (Mieres)	Sociedad Unión Asturiana
1843	Perla	301	El Ruciu (Mieres)	Sociedad Unión Asturiana
1843	Providencia	365	La Flecha (Mieres)	La Fraternidad
1844	La Esperanza	312	El Terronal (Mieres)	José María Cavanilles
1844	Confianza	408	Casallena (Mieres)	José Pineda y Cía.
1844	Peregrina	412	(Mieres)	Gabriel Álvarez y Cía
1844	Eugenia 1st	473	Muñón Cimero (Lena)	Anglo-Asturiana
1844	Eugenia 2nd	476	Maramuñiz (Lena)	Anglo-Asturiana
1844	Eugenia 3rd	–	Piedracea (Lena)	Anglo-Asturiana
1844	Caridad	–	La Granja (Castiello)	Gabriel Cienfuegos
1844	Interesante	–	Muñón Cimero (Lena)	Anglo-Asturiana
1844	Deseada	461	Maramuñiz (Lena)	Unión Astur

). However, it was not until six years later that mercury mining truly began to develop near to Muñón Cimero (Lena), an activity that continued intermittently until the early 1970s.

Table 2. Chronological evolution of mining concessions and operating companies from 1838 to 1974.

Corporate Evolution	Year	Key Events
Asturian Mining Company (Anglo—Asturiana)	1844	Beginning of the mining activities in Lena
Compagnie Minière et Métallurgique des Asturies	1853
Société Houillerè et Métallurgique des Asturies	1861
Sociedad Numa Guilhou	1870	Numa Guilhou acquires the Société Houllière et Métallurgique des Asturies
	1876	Reactivating of the concessions in Muñón Cimero
Fábrica de Mieres	1879	Fábrica de Mieres is established
	1881	The Hg and As concentration system us set
Sectoral crisis	1899
World War I	1914	Increase in the value of Hg
	1915	Mining activities at La Soterraña are halted
Compañía Asturiana de Minas Metalíferas	1930	Leasing of the mercury concessions from Fábrica de Mieres
	1933	Electrification of La Soterraña (Electra de Viesgo)
1934 Revolution	1934	Operational standstill
Compañía El Mercurio Astur	1940
Astur Belga de Minas	1947
Minas La Soterraña SA	1948	Minas de La Soterraña is founded
	1950	Recovery operations of abandoned mine workings
	1952	Extraction of the Ulpine mining shaft
	1954	Minas de La Soterraña acquires the Rosario concesion in La Peña (Mieres)
	1956	Installation of the first rotary kiln, manufactured by Aguinaco
	1958	Kiln firs operational and execution of inclined planes No. 1 and No. 2
	1962	Installation of the second rotary kiln and a skip
	1969	Exploitation below 400 m above sea level at the Ulpino mining shaft
International crisis	1972	Employment adjustment and a significant decrease in production
	1974	Cease of the mercury extraction activities in Asturias
	1989	Partial dismantling of the facilities
Sidefluor SA	1994	The mining concessions of the Group Soterraña finish
	2009	Application of a municipal license for the legalization of other activities
	2010	Cease of the activities
	2013	Demolition works

shows the corporate evolution, including influent events, and key events across all historical evolution of La Soterreaña explotation. The earliest phase of mineral exploitation in the Lena district was led by La Concordia de Mieres (Brañalamosa) and the Asturian Mining Company (Anglo-Asturian; Muñón Cimero). The latter dissolved in 1849, following denial of authorization to continue operations and shareholders’ failure to fully disburse capital. Its assets were acquired in 1853 by a precursor of the Compagnie Minière et Métallurgique des Asturies, which subsequently transferred concessions to the Société Houillère et Métallurgique des Asturies (La Hullera y Metalúrgica de Asturias). Extraction proceeded intermittently owing to challenging geology. Cinnabar remained attractive while mercury prices were high and sales were state-regulated; however, operations at the Eugenia mine ceased due to abundant arsenic sulfide (realgar), which rendered mining and furnace processing hazardous [4].

In 1870, industrialist Numa Guilhou acquired La Hullera y Metalúrgica de Asturias, founding a new enterprise under his name to reinvigorate extraction activities. In 1878, a dedicated processing plant was established to manage ores with elevated arsenic levels. By the following year, the Fábrica de Mieres was formally established under a notarized agreement on 23 March 1879. This company expanded its operations to include not only the metallurgical production of iron and steel but also the extraction of various minerals, including coal, iron, manganese, mercury, calamine, and lead sulfide (Spanish Mining Statistics, 1907) (Table 2).

In the early years of World War I, Fábrica de Mieres sought to boost its operational capacity, temporarily reversing the declining trend in mercury and arsenic production. Despite these efforts, mining activities at La Soterraña Mine were halted in 1915. This cessation persisted for fifteen years until the Compañía Asturiana de Minas Metalíferas, with Les Mines de Cabrales as the majority shareholder, leased mercury concessions from Fábrica de Mieres and resumed operations at Mina Eugenia in Muñón Cimero (Table 2). This period saw a reconstruction of the metallurgical treatment plant and the refurbishment of mining infrastructure.

However, mining activities faced another disruption three years later due to the 1934 Revolution, leading to an operational standstill that lasted until the end of the Spanish Civil War. In 1940, Compañía Mercurio Astur reopened the Muñón Cimero operations, although Fábrica de Mieres retained ownership of the mining concessions. Seven years later, this phase concluded when the constituent companies of Compañía Mercurio Astur merged into the Astur Belga de Minas, marking the end of direct involvement in the Muñón Cimero operations. One year after the dissolution of Compañía Mercurio Astur, Fábrica de Mieres, alongside partners Díaz Marés, Rivera Azpiroz, and Sainz Díaz de Lamadrid, founded the company Minas de la Soterraña S.A. to exploit the Muñón Cimero deposit, initially based on the mercury concessions of Fábrica de Mieres, expanding with La Peña concessions in 1954 and modernizing furnaces and site logistics. Two custom rotary kilns (Aguinaco) were introduced from 1958, and post-1959 economic liberalization enabled substantial investments and deepening to ~340 m. Regional mercury mining peaked during 1960–1971. Subsequent U.S. regulation (USDA, 1971) restricting mercury compounds precipitated a demand collapse, leading to closures in Asturias (1973–1974) and workforce absorption by public entities. Later decades saw dismantling of metallurgical systems; in 1994, concessions at Muñón Cimero and nearby groups were annulled for administrative non-compliance. The site’s facilities were repurposed by Siderflúor S.A. (fluorinated fluxes), including a halted attempt to reuse waste-heap materials for the AP-66. Siderflúor closed in 2010; thereafter the site suffered vandalism and unsanctioned dismantling, with a 2013 environmental sanction and immediate suspension of activities.

Characterization of the site

La Soterraña mine was formed by multiple buildings, including offices, a warehouse or a workshop, among others. Additionally, the ore processing furnaces and the disused adit were also located on the main platform, which covers an area of approximately 18,000 m². The four long chimneys were located on the hillside (Figure 2).

As a result of the mining activities carried out throughout the years, two soil heaps were created around the facilities: the smallest one, with a total area of 4000 m² and a total volume of 13,000 m³; and the biggest one, with an area of 26,000 m² and a volume of 65,000 m³. The first one is situated between the road and the buildings, while the second one is located south of the facilities, in the lower part (Figure 2). This heap contains mine tailings, low-grade ore and metallurgical residues and it presents some stability issues with local landslides due to water erosion of its slopes and its settlement on clayey soils. Because of the material extraction, the original slope is unrecognizable. The majority of the waste consists of slag with variable granulometry, ranging from gravel-sized particles to blocks of 8 or 9 cm, but washing sludge can also be observed.

2.2. Topographic Determinations

• Digital Terrain Model (DTM).

This section of the methodology addresses the topographic calculations conducted across five dates, as detailed in

Table 3. Relation of topographic data.

Topographic Data	Date
Classical topographic data	2008
1st ALS PNOA LiDAR data	2012
Photogrammetric data from UAV	2019
Photogrammetric data from UAV	2020
Photogrammetric data from UAV	2023

. The first topographic survey, carried out in 2008, was performed by the company NortAsistencias e Informes S.L. as part of the project “Topographic Survey in the Soterraña Zone (Lena).” For this survey, a Trimble GNSS dual-frequency GPS with sub-meter accuracy was used. In areas lacking GPS coverage, a topographic radiation method was employed using a Pentax R-115 total station, based on known X, Y, Z coordinate bases.

For the 2012 data, topographic information was sourced from 1st PNOA-LIDAR coverage. ALS data were captured under the framework of the National Plan for Aerial Orthophotographs (PNOA-LiDAR) of the Spanish Ministry of Transport, Mobility and Urban Agenda. The 1st ALS data were collected in 2012. The density of points was 0.5 first returns per m² (0.5 pts m⁻²). Data for 1st ALS was collected with a RIEGL LMS-Q680 sensor, installed on a fixed-wing aerial platform, which operated at 1064 nm, with a pulse repetition frequency of 70 Hz, a scanner of 30°, an average flight height of 1300 m on the GRS80 ellipsoid and with an average overlap of 15%.

For the years 2019, 2020, and 2023, high-precision photogrammetric flights were conducted using a Phantom 4 RTK UAV from SZ DJI Technology Co. Ltd. ^® (Nanshan, Shenzhen, China). This UAV system also integrates a GNSS high-precision RTK positioning module, and the TimeSync system, which gives precise and real-time positioning data for each image, thus optimizing photogrammetric results and providing centimeter-level accuracy without Ground Control Points (GCPs). Due to the complex orography, the photogrammetric survey was carried out with a transversal and longitudinal overlap of 70% and at an average height of 100 m. Data processing was carried out with the photogrammetry and drone mapping software PIX4D Mapper v4.4.12 (Pix4D S.A., Prilly, Switzerland and the steps followed included point cloud generation, 3D model construction, feature extraction, and multispectral band generation. The Phantom 4 RTK employs the DJI FC6310R camera (20 MP, mechanical shutter) with a 1-inch CMOS sensor (effective area ≈ 13.2 × 8.8 mm). The camera is factory-calibrated for lens distortion, with calibration parameters embedded in the image metadata to enable deterministic correction during processing. In practice, the photogrammetric workflow additionally performs on-project self-calibration within the bundle block adjustment to refine intrinsic parameters, thereby minimizing residual distortions and ensuring centimeter-level georeferencing when coupled with RTK image geotagging.

2.

Volume calculations.

The next step involves calculating material volumes. This is achieved using the Cut/Fill tool in ArcGIS 10.3, which operates based on the difference between digital terrain models, identifying areas of material loss and gain. This volume calculation method is commonly associated with the mining industry [5] and is widely used for estimating areas of erosion and sedimentation.

2.3. Calculation of the Mass of Pollutants Released

This methodology section focuses on calculating the quantity of Hg and As released into the environment, specifically considering the previously determined lost volumes and the average concentration of each contaminant in the waste landfill.

The average concentrations of Hg and As were derived from an in-depth literature review, as numerous studies have been conducted on the specific soil heap under study [6,7,8,9,10,11].

This comprehensive review allowed for the selection of representative concentration values, ensuring that the calculated contaminant emissions accurately reflect the site’s potential environmental impact.

The mass of pollutants released is calculated as the amount of As and Hg released into river channels through surface runoff. This is determined as follows (Equations (1) and (2)):

$$\mathrm{S}\mathrm{o}\mathrm{i}\mathrm{l} \mathrm{m}\mathrm{a}\mathrm{s}\mathrm{s} \mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s}\left[\mathrm{M}\mathrm{g}\right]=\mathrm{B}\mathrm{u}\mathrm{l}\mathrm{k} \mathrm{d}\mathrm{e}\mathrm{n}\mathrm{s}\mathrm{i}\mathrm{t}\mathrm{y}\left[\mathrm{g}{\mathrm{c}\mathrm{m}}^{-3}\right]\times \mathrm{S}\mathrm{o}\mathrm{i}\mathrm{l} \mathrm{v}\mathrm{o}\mathrm{l}\mathrm{u}\mathrm{m}\mathrm{e} \mathrm{l}\mathrm{o}\mathrm{s}\mathrm{s}\left[{\mathrm{m}}^3\right]$$

(1)

Soil mass loss (Mg, which is 10³ kg) as a function of bulk density (g, which is 10⁻³ kg, cm⁻³, which is 10⁶ m⁻³) and soil volume loss (m³).

$$Mass contaminant \left[\mathrm{k}\mathrm{g}\right]=Contaminant total concentration \left[\mathrm{m}\mathrm{g}{\mathrm{k}\mathrm{g}}^{-1}\right]\times Soil mass loss \left[\mathrm{M}\mathrm{g}\right]\times {10}^{-3}$$

(2)

Mass of contaminant (kg) as a function of total contaminant concentration (mg·kg⁻¹) and soil mass loss (Mg, which is 10³ kg).

3. Results and Discussion

This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.

3.1. Topographic Determinations

Table 4. Estimation of loss volume, in m³, per period and per year.

Period	Total Loss Volume (m3)	Loss Volume (m3) Year
2008–2012	3337	834
2012–2019	1671	239
2019–2020	108	108
2020–2023	405	135

summarizes the volumes lost from the spoil heap under investigation over the period 2008–2023.

Between 2008 and 2012 the greatest volumetric loss was recorded, amounting to 3337 m³ for the four-year interval—an annual mean of 834 m³, or 55% of the total quantified for the entire fifteen-year study period. Losses then fell sharply: during 2012–2019 the cumulative deficit was 1671 m³, equivalent to 239 m³ yr⁻¹, whilst 2019–2020 exhibited the lowest annual rate, 108 m³ yr⁻¹, confirming the downward trend. From 2020 to 2023 the values stabilized, yielding a mean of 135 m³ yr⁻¹, comparable to the previous triennium.

The two early intervals (2008–2012 and 2012–2019) concentrate the largest losses because the digital terrain models (DTMs) underpinning those calculations derive from markedly different sources. Specifically, the 2008–2012 estimate relies on a conventional field topographic survey with decimetric resolution and sparse point density; such detail is inadequate for volumetric balancing at spoil-heap scale and tends to over-estimate excavated volumes when compared with higher-resolution datasets. Consequently, contaminant quantities for 2008–2012 were not calculated because the underlying DTM lacks sufficient spatial fidelity (see Table 3).

By contrast, the 2012 model originates from the first PNOA LiDAR coverage, which, even at a minimum density of 0.5 pts m⁻², affords superior spatial fidelity. When this LiDAR-based DTM is compared with the 2019 model—generated via UAV-RTK photogrammetry without ground control points [12]—the results converge, demonstrating that two high-resolution, geometrically homogeneous sources yield coherent balances. Discrepancies emerge only when either of these models is confronted with the traditional survey, because the absence of rigorous co-registration between DTMs of different provenance introduces systematic biases in elevation differentials [13].

In summary, the anomalously high volumes calculated for 2008–2012 reflect the metrological limitations of the underlying data rather than genuine erosive processes; when computations are performed using a single acquisition technology with adequate resolution, the resulting figures remain stable and reliable. The same problem occurred in the period 2012–2019, but the degree is less pronounced, showing significantly higher values than when the analysis is carried out with the same method and equipment, as in the case of the periods 2019–2020 and 2020–2023.

Table 4 shows an increase in annualized loss from 108 m³ yr⁻¹ (2019–2020) to 135 m³ yr⁻¹ (2020–2023). This difference is plausibly driven by interannual climatic variability—particularly changes in seasonal rainfall totals, storm intensity/frequency, and antecedent moisture—which modulate runoff, detachment, and sediment transport. A formal attribution would require event-scale hydrometeorological data and is beyond the scope of this study [14,15].

3.2. Calculation of the Mass of Pollutants Released

According to diverse studies made in La Soterraña soil heap, the mean total concentrations of As and Hg are 14,331 and 1390 mg kg⁻¹ (

Table 5. Concentrations obtained in different studies on the main soil heap of La Soterraña.

Study	As (mg kg −1)	Hg (mg kg −1)	Description
A multi-faceted, environmental forensic characterization of a paradigmatic brownfield polluted by hazardous waste containing Hg, As, PAHs and dioxins [6]	15,000	890	Soil heap
Geochemical study of a mining-metallurgy site polluted with As and Hg and the transfer of these contaminants to Equisetum sp [8]	17,193	1950	Soil heap
Arsenic pollution and fractionation in sediments and mine waste samples from different mine sites [9]	10,800	1330	Soil heap
Environmental impact of toxic metals and metalloids from the Muñón Cimero mercury-mining area (Asturias, Spain) [10]	29,214	1036.	Soil heap (mean value)
Correlation between Geochemical and Multispectral Patterns in an Area Severely Contaminated by Former Hg-As Mining [11]	9920	860	Near soil heap
Mean values *	14,331	1390

* Mean value: the average value obtained from all the cited studies. Note: [6] 4 samples at 25 cm of depth, [8] 10 samples at 20 cm of depth, [9] 3 samples at 25 cm of depth, [10] 8 samples at 15 cm of depth, [11] 5 samples at 20 cm of depth.

). In calculating the mean values, only studies conducted within the same spoil heap were considered, and the intra-heap heterogeneity documented in those works was explicitly accounted for in our synthesis.

The Principality of Asturias Soil Screening Levels (SSLs) [16] for human-health protection are 200 mg kg⁻¹ As and 100 mg kg⁻¹ Hg for industrial land use, 40 mg kg⁻¹ As and 10 mg kg⁻¹ Hg for residential/recreational soils, and 40 mg kg⁻¹ As plus 1 mg kg⁻¹ Hg for agricultural–forest soils. Annex III of Royal Decree 9/2005 [17] considers a site contaminated when any concentration equals or exceeds 100 × its SSL. This threshold is surpassed for As under every land use scenario and for Hg under residential and agro-forest end-uses, classifying La Soterraña site as a contaminated location.

In relation with natural background ranges of As and Hg, they are two to five orders of magnitude lower: 0.1–69 mg kg⁻¹ As [18] and 0.03–0.10 mg kg⁻¹ Hg (global mean ≈ 0.06 mg kg⁻¹) [19]. Thus, even the minimum values measured at La Soterraña are 10³–10⁴ times above crustal baselines, placing the spoil heap among the most severely impacted sites reported in Europe.

The environmental and health implications are severe: accidental ingestion or inhalation of dust from the spoil heap would result in chronic intakes that vastly exceed reference values—0.3 µg kg⁻¹ day⁻¹ for inorganic arsenic and 0.2 µg kg⁻¹ day⁻¹ for elemental mercury—posing a direct toxicological threat to exposed populations; furthermore, analyses of near-heap samples demonstrate an off-site contaminant halo capable of degrading surrounding soils and surface waters.

3.2.1. Soil Mass Loss

In the regional context of annual mass losses from mining landforms, Martín Duque et al. [14] show that at Mazarrón (SE Spain), a DEM-of-difference over 1968–2009 yielded 151.8 t ha⁻¹ yr⁻¹, a value reported as typical for actively eroding mined landforms. In the Hiendelaencina district (Central Spain), LiDAR-based volumetrics for 2009–2014 resolved a net loss of 8849 m³ of sludge and a mean erosion rate of 346 ± 9 t ha⁻¹ yr⁻¹, with 10.3 ± 0.6 t of potentially hazardous metals mobilized in six years [15].

Therefore, La Soterraña’s bulk fluxes fall within the range implied by these European tailings sites; for an actively eroding footprint on the order of a few hectares, they correspond to specific losses of ~10²–10³ t ha⁻¹ yr⁻¹ (10–100 kg m⁻² yr⁻¹), underscoring their environmental and geotechnical significance.

According to Menéndez et al. [20], the bulk density of La Soterraña soil heap is 1.67 g cm⁻³. Consequently, Table 5 summarizes the annual soil-mass losses recorded between 2012 and 2023. The greatest depletion occurred during 2012–2019, with an average of 398 Mg yr⁻¹, whereas the lowest was observed in 2019–2020 at 180 Mg yr⁻¹ (

Table 6. Estimation of soil mass loss per year (2012–2023).

Period	Bulk Density (g cm−3)	Soil Mass Loss (Mg)	Soil Mass Loss (Mg) per Year
2012–2019	1.67	2790	398
2019–2020	1.67	180	180
2020–2023	1.67	676	225

). Despite the gap between these extreme values, all rates are alarmingly high. Beyond the volumetric deficit, such losses compromise structural safety: the embankment’s self-weight ordinarily provides basal stabilization, but as material is removed the confining stresses diminish and the stress field—especially at the toe—is redistributed, thereby lowering the overall factor of safety.

Erosion incises gullies and leaves residual slopes steeper than the loose material’s angle of repose, triggering retrogressive failures that can propagate upslope towards the crest. Geometry is decisive: effective height increases or local slope angles exceeding ~25–30° in heterogeneous mine wastes have been repeatedly linked to instabilities [21].

Simultaneously, removal of cohesive fines renders the residue sand-gravel dominated and more porous. The resulting rise in infiltration elevates pore-water pressures and lowers shear-strength parameters (c, φ), further reducing the factor of safety, particularly during intense rainfall events. Internal voiding and erosional undercutting also rob the heap of support, producing local subsidence and surface fissures that jeopardize nearby infrastructure, as evidenced by distress recorded on haul roads adjacent to the dump [22].

Excavation at the toe eliminates the counterweight that resists the main body’s lateral thrusts—a mechanism repeatedly identified as the precursor to catastrophic failures in mining waste landfills [21].

In summary, the substantial mass loss undermines global stability and promotes internal deformations, necessitating prompt geotechnical intervention and continuous instrumental monitoring.

3.2.2. Mass Contaminant

Between 2012 and 2023, the waste heap exported an estimated 52,270 kg of As and 5070 kg of Hg (

Table 7. Estimation of total mass of As and Hg, in kg, between 2012 and 2023.

Period	Total Mass Contaminant (kg)		Mass Contaminant (kg) per Year
	As	Hg	As	Hg
2012–2019	39,992	3879	5713	554
2019–2020	2585	251	2585	251
2020–2023	9693	940	3231	313

). Annualized over the 11-year interval, this corresponds to mean releases of ≈4.75 t yr⁻¹ As and ≈0.46 t yr⁻¹ Hg. The 2012–2019 period dominates the budget, accounting for 39,992 kg As and 3879 kg Hg (≈5.71 t yr⁻¹ As; ≈0.55 t yr⁻¹ Hg), followed by 2020–2023 with 9693 kg As and 940 kg Hg (≈3.23 t yr⁻¹ As; ≈0.31 t yr⁻¹ Hg). The lowest fluxes occurred in 2019–2020 (2585 kg As; 251 kg Hg).

Given the use of site-specific mean total concentrations for As and Hg, these figures represent conservative, mass-based releases to fluvial pathways associated with surface runoff. Uncertainties are primarily driven by (i) spatial variability in waste composition and (ii) residual biases in cross-epoch DTM co-registration; nevertheless, the internal consistency across high-resolution datasets, 2019–2020 and 2020–2023, supports the robustness of the reported loads.

Focused on the values obtained for these periods, 2019–2020 and 2020–2023, the annualized releases (Table 7) are incompatible with health-based and environmental benchmarks. From an ecological-status standpoint, sustained Hg fluxes at the 2020–2023 rate conflict with the Water Framework Directive regime, which designates mercury a priority hazardous substance, sets a biota Environmental Quality Standards (EQS) of 20 µg kg⁻¹ (wet weight) for fish, and requires the progressive reduction and phase-out of emissions [23,24]. Regionally, continued export at ~3.2 t As yr⁻¹ and ~0.31 t Hg yr⁻¹ would also degrade downstream compliance prospects with Water Framework Directive (WFD) objectives and elevate the likelihood of Hg biomagnification in fluvial biota, given EU-reported background medians for Hg in topsoils (median 38.3 µg kg⁻¹ across the EU LUCAS dataset [25]) and the very low EQS for biota [23].

In conclusion, the magnitude of As and Hg released from La Soterraña during 2019–2023 constitutes a critical public-health and environmental risk. The quantities involved (Table 7) overwhelmingly exceed national, European, and international standards designed to protect human health and ecosystems. There is a clear danger of toxic effects via direct contact and via direct and indirect ingestion (through water and food), primarily at the local scale and, to a lesser extent, at the regional scale (transport through the river basin). Given the extreme As and Hg toxicity of—carcinogenicity, bioaccumulation, and multi-organ damage [26,27,28]—the observed concentrations surpass any tolerance thresholds. The current regulatory framework mandates decisive interventions once guideline values are exceeded, and containment and remediation actions have fortunately already begun.

In light of the above, it is necessary to implement in situ measures, including geotechnical stabilization of the ground as a core remediation action, with the aim of preventing the dispersion of contamination into environmental media (soil, surface water, groundwater, and air). Within this framework, environmental rehabilitation works are being initiated—encompassing slope and spoil-heap stabilization, runoff and erosion control, and the sealing or containment of contaminated materials—funded by Spain’s Ministry for the Ecological Transition and the Demographic Challenge (MITECO).

4. Conclusions

This study demonstrates that inter-epoch DTM differencing is a robust approach to quantify soil loss and associated contaminant export from legacy mine waste. Results show that methodological consistency is decisive: epochs generated with UAV RTK photogrammetry (2019–2020 and 2020–2023) yield coherent, stable balances, whereas cross-comparing heterogeneous sources (classical survey vs. LiDAR vs. UAV) introduces systematic co-registration bias that inflates volumetric change. For steep, rough, and compositionally heterogeneous spoil heaps, cm-level RTK georeferencing and a single acquisition–processing workflow across all epochs are essential to minimize vertical error propagation in the DEM. Quality control should include independent checkpoints (RMS_z), and standardized reporting of vertical accuracy and alignment residuals.

High-fidelity large-scale mapping is not optional: it is the precondition for reliable mass budgets. Detailed cartography (≤8–10 cm GSD orthomosaics and DTMs) is needed to resolve rills, gullying, berm degradation, and toe undercutting, as well as hydrological connectivity that governs sediment mobilization. Repeated UAV-RTK acquisitions (seasonal/annual) should be paired with rigorous co-registration and consistent surface filtering to separate ground from waste landfill faces and vegetation. This enables uncertainty-bounded DoDs and defensible conversion from volume loss to soil-mass loss and then to As/Hg loads.

From a risk perspective, the site presents dual criticalities. Geotechnically, gully incision, removal of fines, steep residual slopes (>25–30°), and toe excavation reduce confining stresses, elevate pore pressures, and decrease the factor of safety, favoring retrogressive failures and local collapses that threaten nearby infrastructure (road AS-231) and receptors. Environmentally and in public health terms, the quantified annualized releases in 2019–2020 and 2020–2023 (Table 7) are incompatible with health-based and ecological benchmarks, driving exceedances in receiving waters and sediments, and enabling bioaccumulation/biomagnification (Hg) in fluvial biota. Immediate management priorities are thus to reduce soil loss and hydraulic connectivity (surface water interception, regrading, engineered covers), stabilize slopes (toe buttressing, armoring, check structures), and treat/contact-isolate contaminated runoff, alongside access control (public and livestock) and a multi-media monitoring network (water, sediment, dust, biota).

Limitations and future work. The contaminant-mass estimates rely on mean total concentrations and a single bulk-density value; however, the spoil heap is plausibly heterogeneous in both geochemistry (As, Hg totals; speciation—As(III)/As(V), Hg phases; sulfur content) and edaphic properties (grain-size distribution, fines content, bulk density, TOC, pH, redox, moisture, clay mineralogy). To address this, we will adopt a vertically and laterally stratified design—distinguishing crest, mid-slope, and toe zones—with multiple lateral replicates per zone and depth-resolved increments (surface, shallow subsurface, subsoil), complemented by dedicated bulk-density coring for volume-to-mass conversion. Coupled channel-sediment and suspended-load sampling during storm events will tighten flux estimates and constrain source–sink dynamics. These improvements, integrated with method-consistent UAV-RTK time series and uncertainty-aware DTMs, will reduce parameter uncertainty and deliver regulator-grade mass-release assessments to guide remediation design and compliance tracking. Geostatistical analysis—experimental variography and conditional simulation—will upscale point measurements to continuous concentration fields and propagate spatial uncertainty into site-wide contaminant mass and flux estimates.