Following the Recommendation 2021/2279 provided by the European Union (EU Commission, 2021), environmental impacts have been estimated using the Environmental Footprint EF v3.1 impact assessment method (European Commission and Joint Research Centre, 2022). The EF v3.1 is a methodology that considers 25 significant impact categories including acidification, climate change, land and water use, and material resources (all of which are listed in the first column of Table 4). The E-LCA results, also shown in Table 4, are presented in detail in section 3.2.1.

The social impact assessment was carried out using the Social Impact Weighting Method (Eisfeldt, 2017) provided by the soca v3 module, a PSILCA-based add-on for the ecoinvent database, which provides 83 social impact indicators, related to 18 impact categories and arranged in four different stakeholders (Workers, Local community, Society and Value-chain actors). The impact categories and their related indicators are shown in Tables 5–8 for each stakeholder. The results contained in these tables are analyzed in section 3.3.1.

Working hours, which are related to 1 USD of process output, are the activity variable considered by soca to calculate social impacts. openLCA based on the ratio of unit labor costs to mean hourly labor cost per employee is used for that calculation; the unit labor costs are obtained by dividing the employee compensation by the gross output (both in USD per country, sector and year). All working hours are related to a specific risk level (very low, low, medium, high, very high or no risk) for each social impact indicator. Hence, the results are measured in risk hours. In addition, total social impacts are provided by soca in terms of equivalent medium-risk hours. A factor of 10 is normally applied from one risk level to the next one (e.g., 10 low-risk hours are the same as one equivalent medium risk hour). However, the methodology of Erauskin-Tolosa et al. (2021) is followed in this paper in order to avoid the arbitrariness of these equivalence factors. Each risk level is assigned the central value from the value range considered by PSILCA for each social impact indicator. For example, in the case of forced labor, from 12 to 16 people “trapped in jobs into which they were coerced/deceived and cannot leave” per 1,000 inhabitants is considered a high risk; hence, we assign the high risk a value of 14. This way, by using the number of hours at each risk level and their assigned values, a weighted average of this risk value can be calculated for each indicator. In the case of forced labor, we obtained a risk value of 5.16, which corresponds to a low risk. This value corresponds to what has been called the indirect social impact indicator (Erauskin-Tolosa et al., 2021), which provides information on the socio-economic context of the studied process (i.e., the social footprint). However, the direct impact can be estimated by considering an annual total of 1,800 h per worker. The given example shows a total of 21.42 people in forced labor to produce 473 GWh of RPV electricity as modeled.

Table 4 presents the environmental impacts associated with generating 473 GWh of electricity through RPV deployment in Vitoria-Gasteiz, considering all impact categories defined by the Environmental Footprint (EF) v3.1 impact assessment method. Both absolute values and relative impacts—expressed as a percentage of the impacts from the Spanish conventional electricity mix, which contains only 6.7% of non-open ground PV and 11.2% for total PV (Wernet et al., 2016), for the same amount of energy generated—are provided to facilitate the interpretation of these result. This comparison allows for a clear understanding of the environmental benefits of RPV in contrast with conventional energy sources.

Results confirm that RPV technology has lower impacts across all impact categories when compared to the conventional electricity mix. Metals/minerals material resources, among the most relevant impact categories, presents the highest relative impact, accounting for 93% of the conventional mix impact, primarily due to the significant consumption of metals like gold (which is associated with inverters), silver and tin. The environmental impact of the mining of gold, copper, silver, and tin is summarized in section 3.2.2. This result may seem counterintuitive, since, as shown previously, RPV systems entail substantial primary extraction material demand. However, the high metals/minerals material resources impact in the conventional mix is mainly due to copper mining for distribution infrastructure (including transformation and transport), and while distributed generation—a key advantage of RPV—may reduce such needs, large-scale RPV deployment would still require certain distribution and storage systems, which remain beyond the scope of this study.

In contrast, non-renewable energy resources shows a much lower impact, at only 9% of the conventional mix, highlighting the advantage of photovoltaic systems in reducing fossil fuel dependence. The consumption of fossil fuels, despite being inexistent in the direct PV energy generation, concerns the extraction of raw materials and the manufacture of the different parts of the installations.

With an impact equivalent to 58% of the conventional mix, ecotoxicity is significant too, mainly attributed also to the extraction and processing of raw materials, which is often accompanied by pollution and tailings generation as explained in the review of the following section. In terms of climate change impacts, RPV technology demonstrates significantly lower emissions compared to the conventional mix, with the main contribution coming from fossil emissions (21%). Acidification and eutrophication impacts remain below 30% for all subcategories, with freshwater eutrophication being the most affected (65%), driven by phosphorus emissions related to module production. Land use and water use impacts are relatively noteworthy, at 30 and 37% of the conventional mix, respectively. Human toxicity (40%) and ozone depletion (50%) impacts, although present, remain moderate, and ionizing radiation is instead significantly lower, accounting for only 3% of the impact of the conventional mix.

The seven impact categories marked with an asterisk (*) in Table 4 have been further analyzed to identify their correlation with specific components of the PV system and the geographical location of their occurrence (see Figures 2, 3). These disaggregated analyses enable a deeper understanding of the hotspots within the photovoltaic value chain and potential areas for improvement. Regarding the contribution of the different parts of the RPV installation to the selected impact categories, the environmental impacts corresponding to operating the system are so negligible compared to the other ones that we do no longer refer to them in the environmental LCA.

Figure 2 illustrates the relative contribution of the different components of the PV system to the total environmental impacts. Among these, the PV panel exhibits the highest contribution in most impact categories, particularly in climate change (62%), non-renewable energy resources (69%), land use (82%), and water use (88%). These impacts are mainly linked to the energy-intensive manufacturing of silicon wafers and the extraction of critical raw materials. Silicon itself accounts for 68% of the impact in water use and approximately 30% in both climate change and non-renewable energy resources. Furthermore, about 50% of the land use impact is associated with the production of wood boards, which are used for manufacturing flat pallets for transporting PV panels. This emphasizes the importance of logistics and packaging in the overall environmental footprint.

The inverter also entails a considerable share in metals/minerals material resources (51%) and in freshwater ecotoxicity (30%), reflecting the environmental burden associated with electronic components and metal use. In particular, the gold factory (which according to openLCA includes the extraction and processing of both gold and silver) is a key contributor, accounting directly for 25% of freshwater ecotoxicity and 48% of metals/minerals impacts. Similar to the inverter but on a smaller scale, the electrical installation also contributes to these two impacts, while additionally being notably relevant in terms of acidification (29%). This is primarily due to the extensive use of copper wiring and other conductive materials, which significantly influence the environmental burden of this component. In fact, the copper industry alone contributes to 28% of metals/minerals and 27% of acidification impacts, and is essential in all conductive elements. Finally, the mounting system plays a secondary but still notable role, especially in acidification (15%) and climate change (19%), due to the use of metal structures involving processes with significant emissions.

Figure 3 provides information on the localization of impacts. The results indicate that the greatest part of the environmental impacts occurs outside Europe, especially in the categories of freshwater ecotoxicity (46%) and non-renewable energy resources (68%). In terms of metals/minerals material resources, it is even more striking, as 100% of the geographically defined impact occurs outside Europe. This reveals a strong dependency on imported materials from regions beyond the continent, with significant environmental consequences.

Within Europe, the most notable impact is in the category of land use (67%). As explained above, a substantial portion of this impact (50%) is related to wood forestry for pallet production, principally linked to Sweden and Germany by the ecoinvent database used. Europe also has a considerable share of impacts related to climate change (30%) and water use (64%). The water consumption is down to the use of water resources for silicon processing. Moreover, this industry demands a significant amount of energy, mainly supplied both by hydropower generation—which further increases water consumption—and by fossil fuels, which substantially contribute to the climate change impact. Lastly, despite being the focus of the photovoltaic deployment raised in this study, Spain shows practically negligible environmental impacts in all categories, with values not exceeding 1%.

We can trace the environmental impacts of mineral mining for every time stage of the activity. Before the extraction starts, ecosystems and previous land uses are disrupted—particularly in open-pit mining, but also underground due to infrastructure. Luckeneder et al. (2021) found that 79% of global metal ore extraction in 2019 took place in five of the six most species-rich biomes, and half occurred within 20 km of protected areas. Further, 90% of sites are in areas of below-average water availability, with copper and gold mining often situated in water-scarce zones.

Metals such as copper, gold or silver is frequently extracted from sulfide ores, which also applies to some coalfields. A major issue is acid mine drainage, where sulfides exposed to air oxidize and acidify surrounding waters. This not only lowers pH but also dissolves heavy metals like lead, zinc, or nickel, contaminating ecosystems and communities downstream of a metal/coal mine (Pandey et al., 2007).

The generation of tailings in the world today is immense, since only ounces or pounds of metals are extracted for each ton of ore processed. In 2022, it was estimated that more than 14 billion tons of mining tailings were produced in the world (Cacciuttolo and Cano, 2022). For instance, the Xianjiang area is the most polluted area in China due to soil and water contamination by arsenic, partly natural and partly generated by cooper mining and tailings (Karn, 2015). In Northeast India, acid drainage from coal mining has turned farmland into wasteland (Choudhury et al., 2017).

The Ok Tedi mine is a relevant case. Operating in Papua New Guinea since 1984, it is considered a “slow-motion environmental disaster” (Kirsch, 2014). After an earthquake destroyed the tailings dam, 90 million tons of waste were annually dumped into rivers, affecting agriculture, fisheries, and livelihoods (Skrzypek, 2022). Even if operations ceased today, full recovery would take 150–300 years. Similar issues are seen at the Huanuni tin mine in Bolivia, where untreated tailings have polluted rivers, soils, and farmland for a century. The water below Huanuni is so contaminated it cannot support life or be used for irrigation or drinking (Perreault, 2013). In Chile, small-scale mining of copper, silver, and gold—while limited in output—causes significant pollution due to lack of treatment systems and high exposure to toxic materials (Castro and Sánchez, 2003). Acid drainage can persist for decades or centuries and spread over wide areas. Asia’s biggest copper mine (Malanjkhand, India), active since 1882, still shows contamination in water, sediments, and even in local plants and animals despite modern technologies (Pandey et al., 2007).

Tailings and contaminated water are typically collected in dams, but accidents still occur, with devastating effects. Examples include the Mariana (Brazil, 2015), Mount Polley (Canada, 2014), and Ajka (Hungary, 2010) disasters. In the Boac River (Philippines), a 1996 dam failure destroyed river-based livelihoods. Studies show acid leaching from tailings continues to pollute the area (Roche et al., 2017). Although water dam collapses have declined over recent decades, tailings dam failures—due to poor design, lack of oversight, or cost pressures—have become more severe (Roche et al., 2017). Climate change further increases risks to these structures. If companies are not legally bound to maintain post-closure pollution controls, the burden falls on the public, as seen in Summitville (United States), where cleanup cost taxpayers $150 million (Warhurst and Mitchell, 2000).

Certain inputs in mining are highly hazardous, notably mercury and cyanide used in gold extraction. These pose health and environmental risks during regular use and in case of accidents. Cyanide-contaminated water is deadly for wildlife, especially fish, birds, and bats (Eisler, 2004). In Baia Mare (Romania, 2000), a cyanide spill killed 200 tons of fish and harmed protected species (Eisler, 2004). Mining also causes sedimentation in rivers, severely affecting ecosystems. In Brazil, tin mining—legal and illegal—has altered river profiles, destroyed biodiversity, and created irreparable ecological damage (Andrade, 1999).

Aluminum, important in PV panel production, is environmentally damaging during both bauxite mining and smelting. In Malaysia, bauxite mining has caused dust, water and soil pollution, harmed crops, and increased cancer risk among communities (Lee et al., 2017). Aluminum smelting is highly electricity-intensive and closely linked to dam-building, which causes deforestation, habitat loss, displacement, and greenhouse emissions (Fearnside, 2016).

We can expect that the scale of mining’s social and environmental toll increases, as demand rises and lower-grade deposits are exploited, and impacts correlate with the massive volume of mine waste (Mudd et al., 2020; Prior et al., 2012).

Tables 5–8 show the social impacts and their associated risk level for the production of 473 GWh of PV electricity, modeled in this work for Vitoria-Gasteiz. Each social impact indicator in Tables 5–8 is followed by its reference value and a short definition. Reference values are those considered characteristic for the Basque Country and have been obtained from the EHU Aztarna project (Erauskin-Tolosa et al., 2021), except for the sectoral values, where the Spanish values of the electricity sector have been obtained from soca (Eisfeldt, 2017). Results are compared to the production of 473 GWh of electricity with the current mix, modeled by ecoinvent for the year 2020 (where fossil fuels have a share of 45.44%, renewable energies 43.45% and the remaining 8.11% comes from imported electricity). Each table shows the results for a different stakeholder: Workers (Table 5), Local community (Table 6), Society (Table 7), and Value-chain actors (Table 8). Direct impacts are also shown in brackets for some indicators. For example, according to the definition of the frequency of forced labor impact indicator (Table 5), the results show that there are 5.2 people “trapped in jobs into which they were coerced/deceived and cannot leave” for every 1,000 persons. However, taking into account the total amount of hours at every risk level, and considering that every person works an annual amount of 1,800 h, we calculated that 24.7 people would be in forced labor. Although that is not reflected in the tables, it is noteworthy that the activity hours are almost 40% higher for the current electricity mix (a total of 14.0 million hours), as compared to PV electricity production (8.6 million hours). Therefore, direct impacts can be greater for the current mix electricity production, even when the risk level is lower. This difference could be explained by the fact that PV system is simpler, as it requires less work after its installation. However, it is worth reminding that the distribution and storage stages are not being considered in this study, which could lead to an underestimation of the total activity hours.

As can be observed in Table 5, overall, the risk level increases when compared to the reference values for the Basque Country. Six out of 44 impact indicators for the Workers stakeholder present a high or a very high-risk level, as opposed to the four reference indicators that present a high/very high-risk level. However, the risk for the conventional electricity mix is even higher (seven high/very high-risk level categories). The risk level increases with respect to the reference for 13 indicators (frequency of forced labor, trafficking in persons, minimum wage per month, sector average wage per month, weekly hours per employee, women in the sectoral labor force, gender wage gap, DALYs due to indoor and outdoor air and water pollution, Social Security expenditures, violations of employment laws and regulations, workers affected by natural disasters and right to strike). As further explained below (see Figure 2), most of the working hours related to the PV electricity production throughout its life cycle are held outside Spain and Europe, probably in countries with a laxer labor legislation. The risk level decreases for three indicators (men in the sectoral labor force, non-fatal accidents and violations of mandatory health and safety standards) and it does not change with respect to the reference for the remaining indicators. Nevertheless, the risk does increase in some cases. For example, the risk level is very low for the child labor impact category in both the reference and the PV production scenarios, but when taking into account the whole life cycle (i.e., for the modeled PV production), the value of the total child labor indicator is more than doubled, as compared to the reference. Once again, the socioeconomic context of other countries where most of the labor activities are taking place gains relevance.

Regarding the Local community stakeholder (Table 6), the risk tends to increase. Seven out of 20 impact indicators present a high or very high-risk level. The number of high-risk-level indicators is greater than for the reference value, but also greater than for the conventional electricity (in both cases, three out of 20 indicators present a high/very high-risk level). In comparison to reference values, the risk level increases for six indicators (extraction of biomass related to area, pollution, sanitation coverage, drinking water coverage, unemployment and asylum seekers rate) and decreases for five indicators (level of industrial renewable water use, certified environmental management systems, international migrant stock, net migration rate, and immigration rate). The activities related to PV electricity production result in a high pollution level of the countries where they take place. As it can be seen from the results, they are countries with a scarce sanitation and drinking water coverage. These results match with some of the problems reported in the bibliographic search, such as the health problems derived from mining pollution. The low-risk levels within the category access to material resources is striking. The only indicator which presents a high-risk level is the biomass extraction related to area, which directly matches the dispossession of forest land, reported in the bibliographic search. Even if it still presents a very low risk level, the extraction of fossil fuels and ores increases with respect to the reference value (which is 0), but the extraction of industrial and construction minerals decreases. The reason is that this impact indicator is dominated by the hard coal mining operation, for the production of electricity consumed by the manufacturing of Si. This impact is located mainly in India, which is one of the main coal producers worldwide, but has a very high population. Hence, the risk level for this indicator, measured in annual t per capita, is very low. However, the social impacts of coal mining in India are not negligible at all as explained in section 3.3.2.

As Table 7 shows, Society is the stakeholder where the overall risk level increases most. Nine out of 15 indicators present a high or very high-risk level, a much higher number than for the reference scenario (where only one indicator presents a high-risk level) and also higher than the one for the conventional electricity mix (seven out 15 indicators have a high/very high-risk level). While the risk level of only one indicator decreases (embodied value-added total), the risk level associated to 10 indicators notably increases. The activities related to the PV production are shown not to contribute to the economic development of the countries where they take place. The contribution of the sector to economic development indicator presents a low level of opportunity, but what is especially noteworthy is the high illiteracy level, both for young people and for the total population. The impacts related to the health and safety category are also risky. While the domestic general government expenditure on health is moderate, the expenditure of households (out of pocket) and from external resources (such as international organizations) are high, which show that public health expenditures are not sufficient to cover health issues (Eisfeldt and Ciroth, 2016). Even if it presents a moderate risk level, the life expectancy at birth (67.5 years) is notably lower than the reference value (84 years). The Global Peace Index, which is related to political stability and the risk of conflicts of a certain country (Eisfeldt and Ciroth, 2016), also shows a much higher risk level than the reference. This matches the repression and violence environment reported in the bibliographic search.

Finally, as shown in Table 8, two out of the four indicators related to the Value-chain stakeholder, present a high and very high-risk level. This number is the same for the conventional electricity mix, but it is lower for the reference value (one out of four indicators). While the risk level of two indicators increases (anti-competitive behavior or violation of anti-trust and monopoly legislation and social responsibility along the supply chain), none of the categories presents a lower risk.

A detailed analysis of 14 social impact indicators is presented in Figures 4, 5. The indicators with highest risk levels and highest increases of the risk level with respect to the reference were chosen. Figure 4 shows the relative contribution to the impacts of the different sub-processes involved in the PV electricity production: panel production, inverter production, mounting system, electrical installation and system operation. It should be clarified that soca allocates a certain amount of activity hours to the modeled process in the form of “direct contribution,” which should be interpreted as activity hours associated to the system operation. The hours related to the system operation are not calculated through an inventory of the activity, but through the ratio of unit labor costs to mean hourly labor cost per employee (section 2.3.2). As stated in section 2.1, all sub-processes were modeled with a cradle-to-grave perspective, from the upstream to the end of life.

The first column of the figure is not related to the impacts, but to the activity hours, i.e., it shows the share of each sub-process in the total number of hours involved in the process, regardless of their risk level. 48.2% of the activity hours corresponds to the panel production, 21.9% to the mounting system, 21.5% to the inverter production, 6.2% to the electrical installation and 2.3% to the system operation. 28.5% out of the total activity hours correspond to mining and production of metals. Most of them (20.8% of total activity hours) are related to Al production for the mounting system, followed by PV panel production, where 5.5% of total activity hours is related to metal production. On the other hand, only 3.5% of the activity hours are related to Silicon production, most of them related to the PV panel production.

Returning to Figure 4, the total number of hours of each sub-process was weighted by risk level, as previously performed (Erauskin-Tolosa et al., 2021), in order to account for the contribution of the sub-processes into the impact indicators in the remaining columns. The share of the sub-processes remains similar to what observed in the activity hours in the case of most of the indicators. The main differences arise for unemployment, emigration rate and asylum seekers rate, where system operation gains importance, with a share from 8.9 to 11.9%. Inverter production also increases in importance (28.8–30.7% share), which causes panel production to have a lower share (37.7–42.7%). System operation also shows a raise in importance for the gender wage gap impact (8.5%), where there is a drop in the share of mounting system (12.2%). The opposite occurs with goods produced by forced labor, where the share of system operation is null and the mounting system gains a weight of 29.5%.

Figure 5 shows the location of the impacts related to each indicator: Spain, Europe or outside. In the same way as for Figure 2, the total number of hours of each sub-process was weighted by risk level. The locations “Global” or as “Rest of the world” from the ecoinvent database were considered as not defined. 80.2% of the total activity hours are outside Europe and, with a few exceptions, this trend is more pronounced in the case of the impact indicators. Even though it is not shown in the figure, it is particularly worth noting that most of the high risk and very high-risk hours occur in Global Southern countries such as India, China and African or South American Countries.

The most prominent exceptions are emigration rate and asylum seekers rate. Although impacts related to displacements (high rates of temporary male immigration and accelerated social changes due to immigration) have been reported for the mining regions, results show that only 0.5 and 1.5% are outside Europe, respectively, even if the location of 46.7 and 49.6% of the impacts is not defined. These are followed by gender wage gap, where 36.8% of the impacts occur within Spain and Europe, and by unemployment, where 18.5% of the impacts occur in Spain and Europe. These four indicators are those where direct impact (related to system operation) gains importance. These are the kind of impacts that tend to have the greatest influence on the Global North.

Conversely, 100% of the risks related to goods produced by forced labor occur outside Europe. The highest risk hours are in India, related to gravel, cement and limestone production, and in China, related to hard coal and coke production. This impact is followed by anti-competitive behavior or violation of anti-trust and monopoly legislation, where 98.4% of the impacts are held outside Europe, particularly in India and South Africa, related to the hard coal mine operations. 96.7% of the impacts related to illiteracy are also outside Europe, mostly related to the hard coal mining in India.

From a production chain perspective, social impacts and conflicts can start before extraction itself. For instance, problems may emerge due to the access to resources, such as land or water, fostering processes of speculation and dispossession (Bebbington, 2013). This dispossession can be direct (lands bought at low prices or expropriated) or indirect (pollution). In the case of bauxite mining in Adivasi lands of Eastern India, the major impact claimed by communities are the land use changes and the loss of land-dependent rural livelihoods, forest lands, water resources and indigenous sacred hills (Oskarsson, 2017). Mining is the biggest source of displacement in India after dams, with victims potentially numbered in millions (Oskarsson, 2017). In the Bolivian altiplano, water can no longer be used for irrigation or human and cattle consumption, due to more than 100 years of tin mining (Perreault, 2013). Owen et al. (2022) signal that the majority of the world’s mining projects of the transition minerals are located either on or near Indigenous Peoples’ or Peasant lands with adverse conditions for human rights-compatible permitting, consultation, and consent. This can lead to the exacerbation of social inequalities in already vulnerable locations.

Social impacts also occur during mining operation and are related to the health problems derived from mining pollution. For instance, the notorious spill of 151 kg of liquid mercury in Choropampa, near the Yanacocha goldmine in the Peruvian Andes in the 2,000 s, poisoned over 900 villagers and fostered civil resistance (Arana-Zegarra, 2009). Health impacts are also associated with the daily processes of water pollution. An et al. (2004) report on the relationship between tin mining in Yunnan, arsenic contamination and cancer cases. Some 16 million artisanal miners in more than 70 countries, mostly in the Global South, (producing 12–15% of the gold mined worldwide) use mercury to separate gold from the rest of the ore: the harmful health effects of mercury for artisanal miners have been clearly established (Castro and Sánchez, 2003; Nkuba et al., 2019).

When it comes to labor conditions, the case of the region of Xiantiang is relevant. 95% of solar panels worldwide are made up of polysilicon (Department of Labor, 2021) and 35–45% of global polysilicon production and quartz mining takes place in the Xinjiang Uyghur Autonomous Region (Crawford and Murphy, 2023b). Different organizations denounce forced labor and other human rights violations against workers belonging to minoritized groups in the Uyghur region, among others, in the mining and polysilicon production industry (Crawford and Murphy, 2023a).

Other social impacts are related to high rates of temporary male immigration, alcohol problems (James et al., 2021), increased prostitution and male violence against women and children (Mishra et al., 2024), and local conflicts and repression (Arellano-Yanguas, 2023). Social conflicts can lead to community disintegration, repression and violence (EJAtlas, 2025). In Papua New Guinea, the widespread environmental and social impacts of Panguna copper mine resulted in a civil war (Mudd et al., 2020).

In the case of metal mining in Latin America, Arellano-Yanguas (2023) summarizes the main social impacts as follows: “(i) mismatch between expectations and final results; (ii) inequalities between social groups in access to employment and the benefits of mining activity; (iii) co-optation of authorities and distortion of the local political system through support for pro-company candidates in local and regional elections; (iv) growing dependence of society as a whole on the company; (v) accelerated social changes due to immigration; (vi) alcohol and drug abuse; (vii) increased sexual violence and gender inequality as mining activity is highly masculinized; and (viii) when resources are transferred to local governments, they are often accompanied by increased corruption of authorities and the emergence of criminal groups seeking to benefit from these resources”.

Regarding economic benefits, Fearnside (2016) states that aluminum processing fits into the “resource curse” paradigm: the seeming paradox of countries with the greatest natural resource wealth having the highest incidences of poverty and the lowest indices of social wellbeing. According to his study in Brazil, employment and economic benefits of primary aluminum production is very low in comparison to the large socio-environmental impacts (Fearnside, 2016).

Finally, we want to point out to the social impacts of coal mining, because of its relevance in the energy requirements of the transformation of silicon for panels. We focus on India due to the S-LCA results. The greatest social impact of coal mining in India is displacement (Chikkatur et al., 2009; Paltasingh and Satapathy, 2021; Singh, 2008), which implies landlessness/homelessness/joblessness even if resettlement occur (mainly for indigenous people due to their dependency on land and traditional livelihoods), breakdown of family and community structures, greater class and caste conflicts, risk of marginalization worsening of economic situation due to disruption, and cultural breakdown (Chikkatur et al., 2009; Paltasingh and Satapathy, 2021; Singh, 2008). Forceful eviction or involuntary displacement is a violation of human rights according to Paltasingh and Satapathy (2021). In addition, communities adjacent to coal mines suffer from health problems and agricultural impacts derived from air pollution (PM10), water pollution due to acid drainage, diminished water availability and deforestation (Chikkatur et al., 2009; Mishra, 2017), as is the case in Odisha (Hota and Behera, 2015). Water pollution from nearby coal mining can affect drinking water availability and many human activities such as cooking, fishing or swimming and air pollution fosters health problems such as elevated levels of heavy metals in children, birth defects and miscarriages, cancers, heart and kidney diseases and excess deaths (Morrice and Colagiuri, 2013). Other social impacts are related to the changes in population dynamics (rapid and temporary immigration), increased cost of living or addictions (Singh, 2008). It is estimated that up to 70% of India’s current coal work force is employed informally (Banerjee, 2022), leaving aside all the illegal coal scavenging economy. These tasks are primarily carried out by the members from extremely vulnerable Dalit and Adivasi communities (Chandra et al., 2025). Moreover, and given that land ownership and caste hierarchy is often correlated, women from vulnerable communities are often triple burdened with informal casual labor, illegal coal scavenging and social reproductive work (Chandra et al., 2025). Finally, in different locations around the world, difficulty of communities to capture economic benefits and to generate economic alternatives for the future has also been reported (Lockie et al., 2009).