Artisanal and small-scale gold mining (ASGM) presents a significant environmental challenge globally, particularly in regions like the Tapajós River basin in Brazil, one of the world's most affected areas. This research focuses on the Tapajós River basin due to its high ASGM activity, location within the Amazon rainforest, and the conflict between economic activities, environmental protection, and indigenous rights. The study aims to quantify mercury usage and its environmental impact, along with ASGM's contribution to climate change through energy consumption. This research is crucial because while the use of retorts for mercury recovery has become standard, their effectiveness in reducing environmental mercury release remains unknown. Additionally, mechanization of ASGM through excavators raises concerns about the carbon footprint of gold production. Understanding these factors is essential for developing effective policies to mitigate the environmental impact of ASGM while acknowledging the socio-economic realities of those involved.

Existing literature extensively documents the social and environmental consequences of ASGM, including deforestation, toxic chemical release (particularly mercury), and high climate change impact due to fossil fuel consumption. Studies have investigated social conditions, deforestation rates, and health impacts of mercury in the Brazilian Amazon. However, research on the specific amounts of mercury used and recovered in ASGM is scarce, often limited in sample size and lacking comparability across studies. Similarly, research on the energy intensity and climate impact of ASGM is limited, with most studies focusing on individual mines or specific regions and facing methodological challenges in comparing results. This study addresses this gap by providing comprehensive empirical data on a large number of ASGM sites, including those inaccessible to outsiders.

The study employed a mixed-methods approach involving field surveys conducted between 2018 and 2022 at over 100 sampling points across 50 mining sites in the Tapajós region. Data collection involved *in situ* measurements of mercury use, loss, and recovery using retorts (47 measurements); detailed surveys on energy consumption (34 surveys), including diesel, petrol, and electricity; and interviews and questionnaires. The team used snowball sampling to reach mines that were often informal or illegal and thus inaccessible to outsiders. The study accounted for different mining methods (baixão, dragas, filão) and calculated mass balances for mercury at each site. Fine gold content was determined for different gold products (sponge gold, doré). Energy consumption data were converted to diesel equivalents. The study also considered logistics fuel consumption based on satellite data, georeferenced information, and interview data, estimating average distances between fuel stations, towns, and mining sites. Climate change impact was estimated using the ecoinvent v.3.9.1 database. Data quality control involved comparing findings with literature data, machine datasheets, mass balances, and natural laws; non-sensical values were removed while retaining outliers with plausible explanations. Ethical considerations included obtaining informed consent from all interviewees, adhering to the ethical requirements of Pforzheim University's Ethics Committee, and ensuring the anonymity of participants.

The study found that retorts, simple distillation devices for mercury recovery, are effective and affordable, with 88% of garimpos in the study area using them. The average mercury use was 1.7 kg per kg of gold produced, with an average loss of 0.19 kg per kg of gold. In some cases (dragas), mercury recovery exceeded 100%, likely due to the presence of mercury in existing riverbed sediments from past mining activities. This highlights the ongoing legacy of mercury contamination. Annually, this translates to at least 2.5 tonnes of mercury released in the Tapajós region, despite the use of retorts. Extrapolating these findings to state-level and national gold production estimates suggests far higher mercury emissions. Regarding energy consumption, the study showed that ASGM's climate impact varied by mining technique and mechanization level, ranging from 9,750 to 29,200 kg CO₂eq kg⁻¹ Au. Mechanization, particularly the use of excavators, lowered specific energy intensity but might lead to rebound effects, potentially increasing overall climate impact. Logistics contributed only marginally to the overall climate impact. Compared to large-scale gold mining (LSGM), ASGM's climate impact was comparable, though lower in certain cases due to differing energy sources and efficiencies. Industrial gold recycling offers the lowest carbon footprint.

The findings demonstrate the effectiveness of retorts in reducing mercury emissions, but the remaining levels are still considerable, suggesting the need for additional measures to curb mercury use beyond retort adoption. The results highlight the significant contribution of ASGM to climate change, comparable to that of LSGM, challenging the notion of ASGM as inherently more sustainable. The observation of higher than 100% mercury recovery in some cases underscores the legacy of mercury contamination in the region. Increased mechanization might lead to efficiency gains, but also rebound effects, highlighting the need for holistic approaches. The relatively high climate impact of ASGM, even with retorts, has implications for sustainability certifications, which often focus solely on mercury reduction. The study underscores the need for investment in renewable energy and stricter environmental regulations.

This research provides compelling evidence of the significant environmental impact of ASGM in the Tapajós region, despite the use of retorts for mercury recovery. Brazil must implement its National Action Plan (NAP) under the Minamata Convention to reduce mercury emissions effectively, and better formalize ASGM practices with stronger monitoring. Furthermore, the high climate impact of ASGM necessitates significant investment in renewable energies. Future research should focus on understanding the dynamics of mercury methylation and sedimentation, as well as the potential rebound effects of ASGM mechanization. This study’s findings are relevant for understanding and addressing ASGM's environmental impacts globally.

The study's findings are specific to the Tapajós region, and the situation may differ in other areas. The mass balance approach, while suitable for the context, lacks the precision to pinpoint exact mercury loss at each process step. The study's reliance on self-reported data from garimpeiros introduces potential bias. Extrapolations of mercury emissions to larger scales are based on existing production data, which might contain uncertainties.