Evaluation of the Ecological and Health Risk Associated with Abandoned Tailings Storage Facilities: The Case of Montevecchio Levante (Sardinia, Italy)

Abstract

The environmental legacy of mining operations presents significant challenges in managing impacts on ecosystems, public health, and safety. In Sardinia (Italy), the mining history has left a particularly severe burden of abandoned sites, making remediation a regional priority. To address this issue and to effectively prioritize interventions at abandoned mining sites, a relative risk assessment approach was developed by the Sardinia Regional Administration and the Italian National Institute for Environmental Protection and Research. The aim of this paper is to highlight the results and information obtainable with the above-mentioned approach through its application to a real case: the Montevecchio Levante mining district in southwestern Sardinia. The study provides a detailed identification of the factors underlying the high intervention priority associated with the site under investigation. An analytical quantification of the contribution of the main contaminants to the overall risk was carried out through the calculation of specific risk indices. At the same time, the environmental matrices most involved in the contamination mechanisms were identified. The results indicate that the overall risk is largely driven by the presence of carcinogenic contaminants, with cadmium and lead contributing primarily to the risks associated with surface water and soil, respectively. The findings provide a solid basis for developing targeted strategies to mitigate ecological and public health risks in abandoned mining areas.

1. Introduction

The intense extractive activity that has driven global economic development—significantly increased in recent decades due to growing resource demand—has always been associated with the generation of large volumes of mining waste [1]. To comprehend the global scale of the issue, consider that the worldwide volume of mining waste is estimated at several hundred billion tonnes [2], currently supplemented by an annual production of approximately 15 billion tonnes, 10 times higher than municipal waste generation [3]. Within the European Union (EU) specifically, mining and quarrying activities account for roughly 23% of total waste, second only to the construction sector [4].

Waste materials generated and discarded during extraction and processing operations—typically aimed at recovering a specific metal—are classified into four categories: topsoil, formed by the removal of the uppermost ground layer; waste rocks, consisting of low-grade ore rocks removed during the extraction of the target mineral from the host rock; tailings, produced during mineral processing operations to separate the valuable species from the gangue and characterized by high fine-particle content; and slags, resulting from metallurgical treatment for metal refining [3].

Among the various waste types, tailings represent the primary source of environmental and human health risks, as they can constitute over 90% of extracted material [5] and are characterized by significant metal contents. Globally, between 5 to 7 billion tonnes of tailings are produced annually [6,7]. Chile generates 1.6 million tonnes of tailings per day, South Africa’s production reaches 570 million tonnes per year, and in China, over 4000 hectares are occupied by tailings disposal sites [8]. The recent trend of declining ore grades implies that tailings volumes will continue to rise, thereby posing a significant challenge for the mining industry [9].

The primary environmental risks associated with mining waste, exacerbated by the abandoned state of many mining sites, consist of the potential release of heavy metals into the environment and the structural integrity of containment facilities [10]. From a mineralogical perspective, tailings frequently contain sulfide minerals, particularly pyrite (FeS₂), with less abundant occurrences of arsenopyrite (FeAsS), galena (PbS), chalcopyrite (CuFeS₂), and sphalerite (ZnS). When these minerals oxidize upon exposure to exogenous agents, they trigger water acidification through the process known as acid mine drainage (AMD) [11], widely recognized as the most severe environmental issue linked to mining activities. The water, characterized by low pH values (pH < 5), interacts with tailings and leaches heavy metals. Consequently, acid drainage becomes enriched with high concentrations of metals, including As, Cd, Cr, Cu, Hg, Mn, Ni, Pb, Se, and Zn, which can contaminate soils (causing desertification processes) and subsurface environments, affecting both surface and groundwater [12,13,14,15,16,17]. This contamination can extend several kilometers downstream from the mining site [18]. Furthermore, the flat and large surfaces of tailings’ basins are susceptible to wind erosion, which can disperse particulate matter (PM) heavily contaminated with heavy metals into the atmosphere. This issue is particularly pronounced in regions with arid and windy climates [19].

The above-mentioned environmental concerns are compounded by risks associated with tailings dam failures, often attributable to construction methodologies, which can lead to widespread contaminant release [9,20]. Records indicate that between 1910 and 2020, a total of 336 failure incidents occurred, though this figure may represent an underestimation [20].

These risks are further exacerbated by the historical lack of environmental safeguards at many old tailings settling ponds, which were frequently constructed along watercourses where excess volumes were periodically discharged when capacity was exceeded [6]. Consequently, the most significant environmental risks originate from legacy mining sites dating back to last century’s extractive operations. The number and vastness of the areas affected by mining activity and the presence within them of multiple and heterogeneous hazard sources, requires the ability to rationalize the remediation interventions by optimizing the use of available economic resources. The most suitable tool for pursuing this objective is risk analysis, both relative and site-specific.

The research conducted by the University of Cagliari under the Return project (“Multi-Risk sciEnce for resilienT commUnities undeR a changiNg climate”) aims to assess risks associated with abandoned mining areas and develop technical solutions that simultaneously mitigate hazards while recovering marketable resources, in line with circular economy principles. Specifically, the ongoing activities involve implementing combined processes to extract valuable metals for mineral processing industries and potentially recover lower-value inert materials for construction applications or landscape rehabilitation projects.

In this context, the paper describes the following:

• The Italian approach to environmental risk assessment and management of abandoned mining sites. This methodology employs tiered risk analyses designed to (i) establish intervention priorities among mining areas (large-scale analysis), and (ii) evaluate individual hazard sources within each site (small-scale analysis).
• The Montevecchio Levante mining site (Sardinia, Italy) and its risk-based classification as a case study. This site was selected for experimental activities under the Return project, aimed at simultaneously mitigating environmental risks and recovering valuable resources.
• Risk-assessment algorithms developed by the Sardinia Region and ISPRA (the Italian Institute for Environmental Protection and Research) and their application to the case study, identifying site-specific risk drivers. These include (i) primary/secondary contamination sources, (ii) contaminant typologies, (iii) transport pathways, and (iv) vulnerable receptors that collectively determine the overall risk profile.

The aim of the paper is to highlight the results and information obtainable with the above-mentioned dual-scale approach through its application to a real and highly relevant case.

2. Regulatory Framework

Growing concerns regarding risks associated with abandoned mining sites led the European Union to enact the Directive 2006/21/EC [21], which establishes procedures, measures, and guidelines to prevent or minimize adverse environmental and human health impacts from mining waste management. This directive was transposed into Italian legislation through Legislative Decree No. 117 of 2008 [22], governing mining waste management and addressing three key aspects:

• Fundamental concepts and definitions related to extraction waste.
• Storage facility requirements and specifications.
• Operational protocols for extraction waste management.

The Decree applies to the management of extraction waste within production sites through internal storage structures (e.g., tailings dams, stockpiles, or tailings basins), which are exempt from standard waste landfill regulations. These structures are classified according to their potential risk to human health and the environment, considering the presence of hazardous waste and the likelihood of major incidents (e.g., structural failures).

The required environmental safeguards for these facilities are determined by risk analysis. Rather than setting general technical specifications, the Decree establishes site-specific environmental and safety objectives that take into account local morphological, geological, and environmental aspects.

Article 20 of Legislative Decree 117/2008 mandates the comprehensive inventory of these storage facilities, while Ministerial Decree No. 16 April 2013 [23] establishes the standardized methodology for compiling a national register of decommissioned/abandoned mining waste storage sites. This is achieved through systematic collection and organization of regionally sourced data across Italy.

Another key legislative instrument in the Italian framework is Law No. 98 of 2013 [24], which authorizes the beneficial reuse of mining waste from abandoned sites located in areas of environmental or economic significance. These materials may be repurposed within the same areas for various rehabilitation applications including backfilling, reshaping, embankment construction, land/roadway improvements, and other restoration works, upon demonstrating compliance with contamination standards specified in Italy’s Environmental Code (Legislative Decree No. 152 of 2006 [25]) for polluted sites.

3. Risk Management of Abandoned Mining Sites

Mining areas represent both a territorially significant sector and a substantial environmental challenge. According to a comprehensive survey conducted in 2006 by the Agenzia per la Protezione dell’Ambiente e dei servizi Tecnici (APAT, now ISPRA—Italian National Institute for Environmental Protection), Sardinia—focal region in the framework of the Return project—contains 427 documented abandoned mining sites, with their spatial distribution illustrated in Figure 1.

Given the substantial number of abandoned mining sites, their spatial distribution, and the multiple hazard sources they contain, the Italian regulatory system has mandated the implementation of comparative risk assessment (CRA) methodology. This approach prioritizes sites and their internal hazard sources based on quantified risk indices, to focus the remediation efforts on highest-risk areas first.

This paper presents two distinct CRA frameworks developed for mining waste management in Italy:

• The Sardinia Region’s area-wide assessment protocol for mining districts.
• ISPRA’s site-specific methodology for ranking individual hazard sources within mining areas, particularly tailings storage facilities.

4. Relative Risk Analysis for Mining Sites

For the various mining areas across the region, the Sardinia Regional Government has developed a comparative risk assessment model that assigns each site a standardized hazard score. This methodology is grounded in the identification of environmental sensitivity parameters, detailed in

Table 1. Environmental sensitivity parameters identified by the Sardinia Region for the prioritization of abandoned mining areas. The columns report the range of scores that each parameter can assume, with the description of score_MIN and score_MAX, and the respective weight, based on the potential of that parameter to generate hazards.

Environmental Sensitivity Parameters	Scoring Range	ScoreMIN	ScoreMAX	Weight
Polluted site size	1–4	1 (<1000 m2)	4 (>12,500 m2)	0.75
Embankment methods	0.5–3	0.5 (no waste)	3 (no safeguards)	1.5
Land use in the site’s context	1–5	1 (industrial)	5 (residence)	0.75
Distance to residential areas	0–4	0 (>5000 m)	4 (up to 100 m)	1.5
Type of the primary source of contamination	0.5–4	0.5 (no waste)	4 (special waste)	2
Dimensional characteristics primary source of contamination	1–4	1 (up to 1000 m3)	4 (25,000 m3)	0.75
Environmental matrices affected	2–8	2 (sediments)	4 (multiple matrices)	2
Chemical-toxicological parameters	1–6	1 (no hazardous waste)	6 (suspended products)	2
Groundwater vulnerability	2–6	2 (Low)	6 (very high)	1.25
Groundwater level	0.5–4	0.5 (21–50 m)	4 (2 m)	1.5
Water bodies in the proximities	1–4	1 (>1000 m)	4 (0–100 m)	1.25
Presence of wells and/or springs	0–5	0 (none)	5 (<50 m)	0.75
Wells and/or springs resource use	0–3	0 (none)	3 (public use)	1.25
Natura 2000 network	0–1	0 (absent)	1 (included)	1
State of procedural progress	0.5–5	0.5 (completed)	5 (remediation project)	2
Belonging to a SIN	1–4	1 (no)	4 (yes)	1

, which specifies for each parameter (i) the scoring range, and (ii) its weighted contribution reflecting its influence on contamination risk probability.

The total score P is the hazard score of the site, calculated according to Equation (1):

$$\mathrm{P}={\sum }_{\mathrm{I}{\mathrm{D}}_{\mathrm{F}\mathrm{a}\mathrm{c}\mathrm{t}\mathrm{o}\mathrm{r}}=1}^{\mathrm{n}}(\mathrm{P}\mathrm{u}\mathrm{n}{\mathrm{t}}_{\mathrm{I}\mathrm{D}\mathrm{F}\mathrm{a}\mathrm{c}\mathrm{t}}·\mathrm{P}\mathrm{e}\mathrm{s}{\mathrm{o}}_{\mathrm{I}\mathrm{D}\mathrm{F}\mathrm{a}\mathrm{c}\mathrm{t}}),$$

(1)

where

• ID__Fact specifies the environmental sensivity parameters, ranging from 1 to 16.
• Punt_{ID_Fact} indicates the parameter’s assigned score.
• Peso_{ID_Fact} is the weight assumed by each factor.

The calculated score (P), as described above, is normalized to a decimal scale to enhance numerical interpretation. For each analysis parameter, the observable minimum and maximum scores are considered to establish a numerical range encompassing all total scores determined for the evaluated sites. The normalized total score (P₁₀) is calculated as

$${\mathrm{P}}_{10}={\displaystyle \frac{(\mathrm{P}-{\mathrm{P}}_{\mathrm{m}\mathrm{i}\mathrm{n}})}{{(\mathrm{P}}_{\mathrm{m}\mathrm{a}\mathrm{x}}{-\mathrm{P}}_{\mathrm{m}\mathrm{i}\mathrm{n}})}}·10,$$

(2)

where

• P represents the total weighted score for the observed site.
• P_min indicates theoretical minimum score.
• P_max indicates theoretical maximum score.

The resulting scores are subsequently classified into distinct priority tiers corresponding to different intervention levels, as systematically presented in

Table 2. Priority classes assigned to abandoned mining sites based on the P₁₀ value [27].

Priority Level	Level Description	Interval P10
1	High priority	6.25–10
2	Medium-high priority	4.5–6.25
3	Medium priority	2.5–4.5
4	Low priority	<2.5

.

Following the application of this model to all regional mining sites, the Sardinia Regional environmental Agency (RAS) identified 93 sites requiring medium-priority intervention, 28 sites requiring medium-high-priority, and another 28 sites classified for high-priority intervention measures.

5. Comparative Risk Assessment for Waste Storage Facilities

The risk assessment is applied to containment structures reported in the national inventory, for which ISPRA considers two risk typologies: (1) ecotoxicological/public health risk, and (2) geotechnical/structural risk. The study, as performed so far, focuses exclusively on Ecotoxicological and Public Health Risk (EPHR).

The EPHR assessment is conducted in compliance with the Ministerial Decree of 16 April 2013, employing the ARGIA methodology (Analisi relativa di rischio per la gerarchizzazione dei siti inquinati presenti nell’anagrafe) developed by ISPRA for contaminated sites and adapted to abandoned mining areas.

The computational model evaluates three key factor categories for each assessed contaminated site (or waste containment structure in the mining context): (i) contamination sources, (ii) pollutant transport pathways, and (iii) vulnerable receptors.

The total Risk Index (IRI-Indice di RIschio) for a contaminated site is calculated as the sum of risk indices from all primary sources present at the site, according to Equation (3):

$$\mathrm{I}\mathrm{R}\mathrm{I}=\sum \mathrm{I}\mathrm{R}{\mathrm{I}}_{\mathrm{j}}.$$

(3)

The composite risk index for a primary contamination source (IRI_j) is calculated as the sum of risk indices for all risk-relevant contaminants (IRI_jm) (Equation (4)):

$$\mathrm{I}\mathrm{R}{\mathrm{I}}_{\mathrm{j}}=\sum \mathrm{I}\mathrm{R}{\mathrm{I}}_{\mathrm{j}\mathrm{m}}.$$

(4)

The primary source risk assessment procedure starts with the calculation of the Specific Hazard Coefficient (SHC) for all contaminants exceeding regulatory threshold concentrations. The SHC is computed according to Equation (5):

$$\mathrm{S}\mathrm{H}\mathrm{C}=\mathrm{P}\mathrm{L}·\mathrm{I}\mathrm{H}\mathrm{C},$$

(5)

where IHC denotes the contaminant’s intrinsic hazard coefficient and PL represents the pollutant load, calculated as the product of representative contaminant concentration (CR) and the score for the extent of the contaminated zone (E):

$$\mathrm{P}\mathrm{L}=\mathrm{C}\mathrm{R}·\mathrm{E},$$

(6)

Following this calculation, a screening procedure is applied to select only risk-relevant contaminants. The contaminants are first categorized as either carcinogenic or non-carcinogenic, with subsequent calculation of the ratio between each contaminant’s SHC and the maximum SHC value within its respective category. Indeed, the risk analysis includes an initial screening phase to identify the contaminants that are most likely to contribute significantly to the risk. The screening phase is carried out based on the value of the contaminant-specific hazard coefficient: contaminants with an SHC/SHC_max ratio below the 10% threshold are systematically excluded from further analysis, while those exceeding this threshold progress to risk contribution evaluation. This assessment quantifies contaminant-specific risk through Equation (7), which derives the IRI_jm as the sum of weighted scores for secondary source characteristics (PtS_jm), transport pathway efficiency (PtT_i), and receptor exposure vulnerability (PtR_i):

$$\mathrm{I}\mathrm{R}{\mathrm{I}}_{\mathrm{j}\mathrm{m}}=\sum \mathrm{P}\mathrm{t}{\mathrm{S}}_{\mathrm{j}\mathrm{m}}·\mathrm{P}\mathrm{t}{\mathrm{T}}_{\mathrm{i}}·\mathrm{P}\mathrm{t}{\mathrm{R}}_{\mathrm{i}}.$$

(7)

In Equation (7), the subscript “i” denotes the five environmental compartments considered in the analysis: groundwater, surface water, soil, indoor air, and outdoor air. For each relevant contaminant, the secondary source contribution score is computed as a function of the following parameters:

• Contaminant concentration in the specific environmental matrix (when available).
• Extent of contamination.
• HIC.
• Contaminant partition coefficients among matrices (applied when matrix-specific concentrations are unavailable).
• Contamination containment measures.
• Site accessibility conditions.

The transport pathway score is calculated based on the parameters specified in

Table 3. Parameters considered for determining the contribution to EPHR for each transport pathway. The score range that each parameter can assume is reported, along with the description of the minimum and maximum scores.

Transport Pathway	Parameters	Score Range	ScoreMIN	ScoreMAX
Groundwater	Groundwater depth	0.4–1	0.4 (>100 m)	1 (<2 m)
	Aquifer typology	0–1	0 (absent)	1 (free)
	Aquifer lithotype	0.4–1	0.4 (residual soil)	1 (calcareous crust)
	Distance contaminated area from top of water table	0–1	0 (>30 m)	1 (<2 m)
Surface water	Depth of top from contaminated area	0–1	0 (>30 m)	1 (in contact)
	Hydraulic site localisation	0–1	0 (>200 years)	1 (<50 years)
	Flow rate classes	0.3–1	0.3 (>1000 m3/s)	1 (<10 m3/s)
	Annual average precipitation	0.3–1	0.3 (<500 mm/y)	1 (>800 mm/y)
	Hydrographic density	0.5–1	0.5 (0.001–0.015)	1 (>0.030)
	Run-off	0–1	-	-
	Organic carbon fraction	0.3–1	0.3 (>3%)	1 (<1%)
Indoor air	Depth of top from contaminated area	0–1	0 (>30 m)	1 (in contact)
	Synoptic index	0.5–0.7	0.5 (Rimini weather station)	0.7 (Bologna weather station)
Outdoor air	Depth of top from contaminated area	0–1	0 (>30 m)	1 (in contact)
	Synoptic index	0.5–0.7	0.5 (Rimini weather station)	0.7 (Bologna weather station)
Soil	Depth of top from contaminated area	0–1	0 (>30 m)	1 (in contact)
	Vadose zone lithotype	0.1–1	0.1 (silt)	1 (sand)

, which employs assigned scores derived from guideline recommendations rather than measured field data.

The receptor scoring system incorporates both human populations and sensitive ecological receptors identified within a 5 km radius of the primary contamination source. These scores are computed according to the frameworks presented in

Table 4. Computational framework for human receptor contribution to EPHR for each environmental matrix, based on the distance from the source of contamination and the number of residences/workers present in the area.

Receptors	Distance (m)	Exposure Pathways Score
		Groundwater	Surface Water	Indoor Air	Outdoor Air	Soil
Residences of number N	0–100	N	5·N	87·N	13·N	10−3 N
	101–1000	0.1·N	0.5·N		1.3·N
	1001–3000	0.01·N	0.05·N		0.13·N
	3001–5000	0.001·N	0.005·N		0.013·N
Workers of number N	0–100	N	5·N	30·N	30·N	10−3 N
	101–1000	0.1·N	0.5·N		3·N
	1001–3000	0.01·N	0.05·N		0.3·N
	3001–5000	0.001·N	0.005·N		0.03·N
	1001–3000	0.01·N	0.05·N		0.3·N
	3001–5000	0.001·N	0.005·N		0.03·N

and

Table 5. Computational framework for sensitive ecological receptor contribution to EPHR, based on the distance from the source of contamination and the areal extent of sensitive natural areas.

Distance from Contamination Source	Score
0–100 m	${\sum }_i{A}_i$(sensitive areas 0–100 m)/(3.1 × 104)
101–1000 m	${\sum }_i{A}_i$(sensitive areas 101–1000 m)/(3.1 × 107)
1001–3000 m	${\sum }_i{A}_i$(sensitive areas 1001–3000 m)/(2.5 × 109)
3001–5000 m	${\sum }_i{A}_i$(sensitive areas 3001–5000 m)/(5.0 × 1010)

, accounting for three key variables: (i) human receptor population density, (ii) areal extent of natural receptors, and (iii) their respective distances from the contamination source.

Following score computation for all parameters, Equation (7) is applied to each relevant contaminant. The contaminant-specific scores are then aggregated via Equation (4) to determine the risk index for the primary contamination source under investigation. This analytical process is repeated for all primary sources present within the containment structure.

Using the described methodology, storage facilities were classified by ISPRA into five risk categories: low (L), medium-low (ML), medium (M), medium-high (MH), and high (H). The classification employed these normalized cutoff thresholds: L < 43 ≤ ML < 58 ≤ M ≤ 68 < MH < 86 ≤ H.

Updated 2022 data reveal the following distribution among Sardinia’s 209 inventoried storage facilities: 73 with medium IRI, 80 with medium-high IRI, and 56 with high IRI.

6. The Montevecchio Levante Mining Complex

The study area comprises the Montevecchio Levante mining complex, extending across 60 km² in southwestern Sardinia (Figure 2). This site hosted one of Europe’s most productive lead–zinc vein systems.

Mining operations generated approximately 4.3 million m³ of waste, deposited in a tailings basin that was systematically breached during active operations, releasing contaminated sediments into the nearby Sitzerri River system. These materials were subsequently transported and redeposited over a 16 km downstream area, forming extensive secondary contamination zones.

The accumulation of mineral processing fines in the Rio Sitzerri floodplain has caused significant soil degradation, including agricultural land desertification downstream of the tailings basin and disruption of farming/livestock activities. Surface contamination also poses atmospheric pollution risks, particularly during drought periods, through wind-driven erosion of exposed surfaces. The area further exhibits surface and groundwater contamination (notably by Cd, Ni, Pb, Zn, Mn, and Fe), primarily from AMD originating in mine tunnels and the tailings basin.

The principal risk sources include (i) natural mineralized zones intersected by anthropogenic excavations, (ii) the main tailings basin, and (iii) fine-grained waste deposits along the Rio Sitzerri floodplain.

The tailings basin was comprehensively characterized during preparation of the abandoned mine site characterization plan, with 26 boreholes sampling the full deposit thickness across the entire 25-hectare hazard area.

Table 6. Exceedances of inorganic contaminant reference values in soil and subsurface environmental matrices for all 26 sampling points of the Montevecchio Levante tailings basin. Concentration values exceeding CTC_A and CTC_B are indicated in blue and red, respectively.

ID	As	Be	Cd	Co	Cu	Hg	Pb	Sb	V	Zn
	mg/kg
SDBF 01	92	3.30	25	51	79	1.31	672	9	36	3781
SDBF 02	195	0.12	40	38	140	3.70	2512	59	24	6848
SDBF 03	131	0.10	91	52	162	2.35	1475	33	22	10,538
SDBF 04	27	1.81	30	38	98	1.06	1581	22	84	3595
SDBF 05	166	0.10	15	52	450	2.13	20,016	144	51	2661
SDBF 06	229	1.91	59	83	450	2.71	7390	118	35	8861
SDBF 07	101	0.10	62	36	175	1.88	1851	32	28	9357
SDBF 08	109	1.22	55	44	152	2.51	1512	36	30	9622
SDBF 09	96	1.32	42	46	153	1.66	1579	31	33	6184
SDBF 10	92	1.30	45	32	142	2.07	1491	11	25	6518
SDBF 11	84	0.10	54	47	138	1.31	1565	23	28	7177
SDBF 12	124	2.04	43	46	148	1.40	2436	49	53	7081
SDBF 13	112	0.10	32	43	100	1.61	1791	38	33	5137
SDBF 14	145	0.20	16	33	160	1.78	6422	63	24	3556
SDBF 15	111	0.10	46	35	150	1.56	1538	31	26	5643
SDBF 16	76	1.96	62	25	197	1.11	3860	25	36	8563
SDBF 17	105	2.43	85	57	233	0.69	2327	33	50	9856
SDBF 18	112	0.57	32	31	166	1.21	1440	27	31	5186
SDBF 19	139	2.58	80	31	340	1.96	2459	49	58	12,507
SDBF 20	115	0.10	68	53	298	2.80	2464	35	28	9604
SDBF 21	109	1.49	33	36	92	1.01	1822	39	33	5660
SDBF 22	115	0.33	41	25	131	1.77	2185	43	42	6901
SDBF 23	81	1.04	94	50	120	0.54	1183	48	33	10,128
SDBF 24	119	2.13	46	41	153	0.89	3130	43	58	6594
SDBF 25	118	0.10	25	50	119	1.57	1184	26	26	4096
SDBF 26	7	0.48	3	53	23	1.75	392	2	135	347
CTC_A	20	2	2	20	120	1	100	10	90	150
CTC_B	50	10	15	250	600	5	1000	30	250	1500

Analytical results demonstrate that Co, Cu, and Hg consistently surpass CTC_A thresholds (marked in blue) at nearly all sampling locations, while As, Cd, Pb, Sb, and Zn exhibit basin-wide exceedances of both CTC_A and CTC_B limits (indicated in red).

displays average contaminant concentrations from the drilling data, specifically highlighting substances that exceeded at least one Italian regulatory threshold for either public green space soils (CTC_A) or industrial-use soils (CTC_B) as established in Table 1 of Annex 5 to Part IV of Legislative Decree 152/2006.

Among the 26 boreholes drilled in the tailings basin, nine were instrumented as piezometers for groundwater sampling and contaminant analysis. Analytical results were compared against groundwater CTCs (CTC_W), revealing exceedances for Cd, Mn, Ni, and Zn across all monitoring points, with nearly universal violations for sulfates, Pb, and Co. Surface water sampling conducted near the basin perimeter demonstrated contamination throughout the Montevecchio-Sitzerri stream system, Pb, Zn, and Cd concentrations exceeding regulatory thresholds.

The solid-phase, groundwater, and surface water concentration data serve as input parameters for comparative risk assessment using the ARGIA methodology. The following sections detail the application of the risk prioritization frameworks described in Section 4 and Section 5 to the Montevecchio Levante case study, conducting a critical analysis of the key parameters and factors driving the elevated risk associated with both the mining area and its containment structures.

6.1. Application of Comparative Risk Assessment to Mining Sites

The risk assessment input data were derived from technical literature sources and spatial analyses performed using QGIS, software version 3.22.4.

Table 7. Application of Comparative Risk Assessment (RAS methodology) to the Montevecchio Levante mining complex. The columns report the environmental sensitivity parameter, its description, the assigned score (P), the weight (W), and the weighted score (P × W).

Environmental Sensitivity Parameters	Factor Description	P	W	P × W
Size of the contaminated site	60,000 km2	4	0.75	3
Embankment modalities	Waste discarded without environmental controls	3	1.5	4.5
Land use in the site’s context	Forests and natural areas	2	0.75	1.5
Distance from residential areas	<1000 m (town centre Montevecchio)	3	1.5	4.5
Nature of primary source of contamination	Special waste	4	2	8
Dimensional characteristics primary source of contamination	4,300,000 m3	2	0.75	1.5
Environmental matrices affected	Multiple matrices	8	2	16
Chemical-toxicological parameters	Presence of Category 1A carcinogens	5	2	10
Aquifer vulnerability assessment	High	4	1.25	5
Groundwater depth	13 m	1	1.5	1.5
Nearby water bodies	On site	4	1.25	5
Presence of wells and/or springs	Distance between 500 m and 1000 m (Funtana Lacu)	1	0.75	0.75
Wells and/or springs resource use	Drinking use	3	1.25	3.75
Natura 2000 network	Included	1	1	1
State of procedural progress	Characterisation plan	3	2	6
SIN	Belonging to SIN	4	1	4
P				76
P10				7.4

details the characteristics, assigned scores, and weighting factors for all evaluated parameters.

It is highlighted that eight of the sixteen parameters in the calculation framework reflect the most severe contamination conditions documented by ARPAS (Agenzia Regionale per la Protezione dell’Ambiente—Sardinia Regional Protection Agency) at Montevecchio Levante, consequently receiving the highest possible scores.

Analysis of weighted parameter scores reveals that the total risk index is predominantly influenced by (i) multi-media contamination (affecting more than one environmental compartment), and (ii) waste characteristics, particularly the presence of Class A-1 carcinogens (highly toxic substances).

The aggregated weighted score of 76, when processed through Equation (2), yields a normalized P₁₀ value of 7.4. This confirms the site’s high remediation priority, exceeding ARPAS’s intervention threshold (P₁₀ > 6.25) for abandoned mining areas.

6.2. Application of Comparative Risk Assessment to Tailings Storage Facilities

6.2.1. Identification of Priority Pollutants

The comparative risk analysis methodology proposed by ISPRA was applied to the Montevecchio Levante tailings basin structure, identified as the primary contamination source. The analysis included all contaminants from Table 6 exhibiting at least one exceedance of the CTC. These contaminants were categorized as carcinogenic or non-carcinogenic as shown in

Table 8. Calculation framework for the Specific Hazard Coefficient (SHC) of individual contaminants for relevant contaminant selection purposes. (CR—representative concentration; E—extent of the contaminated zone; PL—pollutant load; IHC—intrinsic hazard coefficient; SHC—specific hazard coefficient; σ—SHC/SHC_max).

Contaminant	Toxicity	CR	E	PL	IHC	SHC	σ
As	Carcinogenic	195.51	0.005	0.978	2.6 × 106	2.5 × 106	100
Cd	Carcinogenic	92.42	0.005	0.462	1.1 × 106	5.1 × 105	28
Be	Carcinogenic	2.58	0.005	0.013	1.5 × 106	1.9 × 104	0.78
Pb	Carcinogenic	7148	0.005	35.740	7.6 × 103	2.7 × 105	11
Co	Non-carcinogenic	55.71	0.005	0.279	3.7 × 10−1	1.1 × 10−1	0.76
Cu	Non-carcinogenic	422.39	0.005	2.112	6.2 × 10−2	1.3 × 10−1	0.97
Hg	Non-carcinogenic	2.78	0.005	0.014	2.0 × 101	2.8 × 10−1	2.06
Sb	Non-carcinogenic	104	0.005	0.520	2.6 × 101	1.4 × 101	100
Sn	Non-carcinogenic	5.10	0.005	0.026	6.7 × 100	1.7 × 10−1	1.26
V	Non-carcinogenic	77.30	0.005	0.387	7.9 × 100	3.1 × 100	22
Zn	Non-carcinogenic	10,435	0.005	52.178	2.4 × 10−2	1.3 × 100	10

, which details the following for each: the representative concentration (CR, 95th percentile of Table 6 data), source extension (E), pollutant load (PL), intrinsic hazard coefficient (IHC), specific hazard coefficient (SHC), and the ratio σ (SHC/SHC_max per category).

Among carcinogenic contaminants, As exhibited the highest SHC value (SHC_max), which served as the reference value for relevant contaminant selection. The calculated SHC values for Cd and Pb were 28% and 11% of SHC_max, respectively. Be showed an SHC below the 10% threshold (0.78%) and was consequently excluded from the relative risk analysis.

For non-carcinogenic contaminants, Sb displayed the highest SHC value. Comparative analysis revealed that only V and Zn exceeded the 10% SHC_max threshold among non-carcinogenic pollutants.

Based on the above findings, the risk assessment was conducted with regard to contamination caused by As, Cd, Pb, Sb, V, and Zn.

6.2.2. Source Contribution

For each relevant contaminant, secondary source contributions, transport pathways, and receptor exposures were quantitatively assessed. Based on measured concentrations in soil, groundwater, and surface water matrices along with contamination extent data,

Table 9. Assigned scores for cadmium (Cd) source contribution calculation, according to ARGIA methodology for the 5 environmental matrices.

	Groundwater	Surface Waters		Indoor Air	Outdoor Air	Soil
Concentration	0.526	3.84		92.42	92.42	92.42
Extension	5.0 × 103	5.0 × 103		5.0 × 103	5.0 × 103	5.0 × 103
SHC	5.1 × 105	5.1 × 105		5.1 × 105	5.1 × 105	5.1 × 105
Partition coefficient	1	1		0	6.9 × 10−12	1
Modes of containment	0.9	0.9		0.9	0.9	0.9
Accessibility	-	Residents	Worker	-	1.5 × 10−6	Residents	Worker
		0.5	1			0.5	1
TOT Score	1203	4392	8784	0	0	105,701	211,401

exemplifies the secondary source contribution calculation for cadmium contamination risk. In the absence of direct atmospheric source data, the contribution was derived by multiplying primary source (soil) concentrations by the standardized soil–air partition coefficient.

The containment score assigned to contaminant control measures (row 5 of Table 9) was consistently 0.9 across all environmental matrices, reflecting an open-air stockpile configuration lacking both basal impermeable layers and surface covers. Due to the absence of perimeter fencing and access control systems at the primary source, maximum accessibility scores (Table 9, row 5) were assigned—1.0 for residents and 0.5 for workers. The aggregate risk contribution for each affected environmental matrix was calculated as the multiplicative product of individual parameter scores listed in the corresponding columns.

Table 10. Source contribution for each relevant contaminant across the 5 intervention domains.

	Groundwater	Surface Waters		Indoor Air	Outdoor Air	Soil
		Residents	Worker			Residents	Worker
Cd	1203	4392	8784	0	0	105,701	211,401
As	1555	126	252	0	0	1,118,057	2,236,113
Pb	686	1186	2371	0	0	4,368,529	8,737,058
Sb	0.0004	0.01	0.02	0	0	3	6
V	0.003	0.001	0.003	0	0	0.53	1
Zn	1	1.97	3.94	0	0	29	59

summarizes the secondary source risk contributions for all relevant contaminants, demonstrating significantly greater impacts from carcinogenic contaminants compared to non-carcinogens. Cd exhibited particularly elevated risk scores in surface water matrices, while As and Pb dominated soil-mediated risks. Zn emerged as the most consequential non-carcinogen, though its contribution remained three orders of magnitude below carcinogenic contaminants. The analysis confirmed negligible air pathway contributions due to soil/air partition coefficients approaching zero for all target substances, effectively eliminating indoor and outdoor air matrices as exposure routes.

6.2.3. Transport Pathway Contribution

The environmental risk contribution from transport pathways (represented by different environmental matrices) was quantified using the scoring system presented in

Table 11. Scores assigned to transport pathways representative of the 5 domains. The columns report the transport pathways, the parameter considered, its description, the range score, and the assigned score. The contribution of each transport pathway is calculated as the product of individual parameter scores.

Transport Pathway	Parameter	Description	Range	Score	Contribution
Groundwater	Depth groundwater	13 m	0.4–1	0.7	0.42
	Aquifer typology	Free	0–1	1
	Aquifer lithotype	Sand/silt	0.4–1	0.6
	Distance from groundwater	Direct contact	0–1	1
Superficial water	Depth of top from contaminated area	Superficial	0–1	1	0.17
	Location of the site by hydraulic risk	High hydraulic hazard (Tr ≤ 50 years)	0–1	1
	Flow rate of a surface watercourse	<10 m3/s	0.3–1	1
	Rainfall	727 mm/y	0.3–1	0.8
	Hydrographic density	Assumed	0.5–1	0.6
	Run-off	From nomogram, rainfall function and soil type	0–1	0.5
	Organic carbon fraction	Absence of organic contaminants	0.3–1	1
Indoor air	Depth of top from contaminated area	Superficial	0–1	1	0.70
	Synoptic index	Assumed	0.5–0.7	0.7
Outdoor air	Depth of top from contaminated area	Superficial	0–1	1	0.70
	Synoptic index	Assumed	0.5–0.7	0.7
Soil	Depth of top from contaminated area	Superficial	0–1	1	0.90
	Lithotype of the vadose zone	Silty sand	0.1–1	0.9

. Analysis of the assigned scores reveals significant environmental risk across all transport pathways, with multiple parameters reaching or approaching the maximum score value (1). The composite score for each matrix, calculated as the product of individual parameter scores, yielded the following risk contributions: groundwater 0.42, surface water 0.17, indoor/outdoor air 0.70, and soil 0.92.

For groundwater, a score of 1 was assigned for both aquifer type (unconfined) and source-to-aquifer distance (source in direct contact with the aquifer). The other two parameters influencing groundwater contribution—aquifer lithology and water table depth—received scores of 0.6 and 0.7 respectively, indicating moderate risk contribution.

Regarding surface waters, the flow rate of the Montevecchio stream measured near the Levante basin was 38 L/s, which was assigned a score of 1 as it indicates a low contaminant dilution capacity. Furthermore, the risk contribution from surface waters is significantly influenced by the fact that part of the study area falls within a zone of high hydraulic hazard and contamination is present at the surface. The presence of contaminants at the surface resulted in the maximum score of 1 being assigned to the “depth to the top of the contaminated zone” parameter, which also strongly contributes to the risk associated with outdoor air, indoor air, and soil transport pathway.

6.2.4. Receptors Contribution

The risk contribution associated with the presence of human receptors and sensitive natural receptors located within 5 km of the contaminated site’s perimeter was evaluated by defining four concentric circular zones, as illustrated in Figure 3. These zones correspond to areas lying within 100 m (orange line), 1000 m (yellow line), 3000 m (green line), and 5000 m (blue line) of the site boundary.

The two closest zones (0–100 m and 101–1000 m) contain no residential areas but exclusively occupational buildings, with an estimated average presence of 20 and 30 workers, respectively. The 1001–3000 m buffer zone encompasses the historic village of Montevecchio, with a population of approximately 355 inhabitants. The outermost zone (3001–5000 m) includes the towns of Guspini and Arbus, totaling around 16,900 residents. For computational simplicity, the number of workers in the two outer zones was incorporated into the resident population figures.

The above data were used to compile

Table 12. Risk contribution for human receptors (residents and workers) across the five exposure pathways of interest, based on the distance from the source and the number of residents and workers.

Receptors	Distance (m)	N	Exposure Pathway Weighting Score
			Groundwater	Surface Waters	Indoor Air	Outdoor Air	Soil
Residents of number N	0–100	0	0	0	0	0	0
	101–1000	0	0	0		0
	1001–3000	355	3.5	17.5		45.5
	3001–5000	16,900	16.9	84.5		219.7
Workers of number N	0–100	20	20	100	600	600	0.02
	101–1000	30	3	15		90
	1001–3000
	3001–5000
Contribution			43	217	600	955	0.02

and quantify the risk contribution from human receptors. The total receptor contribution is calculated as the sum of individual scores assigned to the five assessment categories. Notably, the highest risk contribution arises from airborne exposure for workers located in the zone closest to the contamination source, with indoor and outdoor air exposure scores of 600 and 955, respectively.

A significant risk contribution was calculated for both groundwater (score of 43) and surface water (score of 217), while the soil matrix contribution was found to be negligible.

Regarding risks to sensitive natural targets, the highest contribution comes from receptors within the buffer zone closest to the source perimeter, as the entire 0–100 m area is designated as a Site of Community Importance (SCI). As shown in

Table 13. Risk scores assigned to ecological receptors as a function of distance from the contamination source.

Distance (m)	Score
0–100	1.00
101–1000	0.07
1001–3000	0.01
3001–5000	0.00
TOT	1.08

, this results in a score of 1, calculated using the formula provided in Table 4. While significant portions of the outer buffer zones also contain areas classified as SCIs or protected zones under the Regional Forestry Agency (Forestas), their risk contribution remains negligible. This occurs because the calculation formulas in Table 4 incorporate correction factors that make scores highly dependent on the proximity of sensitive areas to the contamination source.

7. Discussion

Several studies in the literature [28,29,30] have highlighted high levels of heavy metal contamination in soil and water in the Montevecchio area. In the present work, two analytical methodologies were applied and their results were analyzed to better understand the causes of the high risk level associated with the study area.

Application on the large scale of the ARPAS method to the entire Montevecchio Levante area provided valuable general information on the following factors, highlighting the primary causes of the elevated risk level:

• Primary contamination source consisting of special/hazardous waste.
• Secondary contamination affecting multiple environmental media.
• Contamination by Class A-1 carcinogenic substances (highly toxic).

In contrast, the site-specific application of the ISPRA model to the deposit area provided detailed insights into the specific factors responsible for the elevated risk levels at the study site.

Table 14. Risk Index (IRI) values for human and ecological receptors, evaluated for both carcinogenic and non-carcinogenic contaminants.

Risk Index (IRI)	Score
Carcinogenic human receptor	9.8 × 105
Carcinogenic natural receptors	1.6 × 107
Non-carcinogenic human receptors	0.23 × 102
Non-carcinogenic natural receptors	0.96 × 102
Total IRI	1.7 × 107

shows the calculated Risk Index (IRI) values for both carcinogenic and non-carcinogenic contaminants affecting human and natural receptors.

The data analysis demonstrates that carcinogenic contaminants dominate the overall risk contribution, with risk scores up to five orders of magnitude higher than those of non-carcinogenic contaminants. Additionally, for both contaminant types, the IRI values for natural receptors consistently exceed those for human receptors. This observed pattern stems from two key factors: first, the risk calculation algorithm heavily weights the proximity of receptors to the contamination source; second, the primary contamination source is situated within a designated SCI—a protected zone for sensitive natural habitats. Notably, the nearest human settlements are located over 3 km from the source, resulting in significantly lower risk scores for human receptors compared to ecological ones.

Figure 4 presents the risk score distribution across the five environmental matrices for both human and natural receptors. With respect to the index risk IRI for human receptors, surface water represents the matrix with the highest contribution, accounting for 64% of total risk. Soil and groundwater account for 30% and 6% of total risk, respectively. Previous studies [31] have reported that heavy metal contamination in surface waters in the abandoned mining areas of southwestern Sardinia was a central issue.

The difference between the contributions of soil and surface water matrices was analyzed based on the risk contribution results related to secondary sources, pollutants, transport pathways, and receptors, as reported in Table 10, Table 11 and Table 12. In particular, the soil matrix was associated with the highest scores for source contribution and transport pathways (Table 11), whereas its contribution to human receptor risk was close to zero (Table 12).

In contrast, surface water showed lower risk contributions from secondary sources and transport pathways, but significantly higher contributions for receptors, resulting in the greatest overall risk contribution. This outcome stems from the calculation framework described in Table 4, which assigns higher scores to exposure pathways that are more effective in conveying contamination to human receptors.

Both indoor and outdoor air pathways demonstrated negligible risk contributions, reflecting the absence of secondary contamination data for atmospheric matrices.

For sensitive natural receptors, soil emerged as the dominant risk matrix. This is due to the fact that, as shown in Table 12, the calculation of risk for natural receptors does not differentiate between exposure pathways across environmental media (Table 13). As a result, the higher source and transport scores associated with soil (Table 11) have more direct and significant impact on the total risk for natural receptors, unlike in the case of human receptors.

Figure 5 presents the aggregated risk contributions of various contaminants through stacked column visualization of IRI values for both human and natural receptors, clearly demonstrating negligible input from non-carcinogenic contaminants (Sb, Zn and V) across both receptor categories. Considering the risk associated with carcinogenic contaminants for human receptors, Cd and Pb account for 52% and 38% respectively, with As contributing 10%. For natural receptors, Pb makes the largest contribution (about 78%). It can be seen, compared to the IRI for humans, that the contribution of As increases to 20% and that of Cd decreases significantly (from 52% to 2%).

The different impact of contaminants for human and natural receptors is explained by the incidences they have in the risk-relevant matrices, illustrated in Figure 6 where the distribution, in percentage, of the risk indices (IRI) calculated for Cd, As and Pb, is shown. From Figure 6, Cd has an impact of almost 80% on the IRI of surface waters, a matrix already identified as the one that contributes most to the risk for human receptors. The results in Figure 6 are consistent with the studies by [32] which found that, in the surface waters of the Montevecchio area, under near-neutral conditions (pH around 6), Cd exhibits higher mobility in water compared to Pb. The latter impacts 78% on the IRI of the soil matrix, which is the most important for natural receptors.

It is important to emphasize that the aim of this study is not the precise quantification of risk indices for the contaminated site. This approach differs from several studies in the literature [33,34,35], which apply absolute risk assessment methodologies and compare the results with specific reference standards, such as those provided by the U.S. EPA for health risk [36]. In this study, instead, the focus was placed on identifying the factors that most significantly influence the risk associated with the contaminated site. This methodological choice made it possible to deepen the understanding of the processes that determine risk and to define operational priorities for its management, while reducing the possibility of directly comparing the results with commonly used threshold values or regulatory criteria.

8. Conclusions

This study addresses key environmental challenges and risk assessment methodologies pertinent to abandoned mining sites. The risk analysis frameworks proposed by the Sardinia Regional Administration (RAS) and the Italian National Institute for Environmental Protection and Research (ISPRA) were applied to the Montevecchio Levante mining area (Sardinia, Italy) and its associated waste deposit facility, respectively, to identify the factors driving both the elevated ecological–health risks and the high intervention priority. Implementation of these methodologies yielded detailed characterization of main contamination sources and their impacts on human and natural receptors.

The results demonstrate that the overall risk is predominantly attributable to carcinogenic contaminants, with particularly elevated concentrations of Cd and Pb, which primarily contribute to surface water and soil-related risks, respectively. Notably, surface water was identified as the main exposure pathway for humans, while soil emerged as the principal risk vector for natural receptors, consistent with the area’s ecologically sensitive context.

These findings provide critical insights into environmental risks at abandoned mining sites and establish an essential foundation for planning remediation strategies.

Consistently, the activities of the University of Cagliari within the Return project are continuing with the experimental phases involving the representative collection of tailings samples from the Montevecchio Levante site and their treatment through a combination of chemical–physical processes including flotation, separation by size and density, green leaching. The combination of treatments is aimed at recovering a Zn concentrate with characteristics suitable for marketing in the metallurgical sector and granular aggregates sufficiently depleted in terms of residual presence of Zn and other associated metals (Cd, Pb, As, etc.) to be reused for landscape fillings and modeling.