Mercury emissions from artisanal and small-scale gold mining across the southern hemisphere surpass coal combustion as the world’s largest source of mercury.We examine mercury deposition and storage in the Peruvian Amazon, heavily affected by artisanal gold mining.Intact forests in the Peruvian Amazon near gold mines received extremely high mercury inputs, with elevated total and methylmercury in the atmosphere, canopy leaves, and soil.Here, we show for the first time that intact forest canopies near artisanal gold mines intercept large amounts of particulate and gaseous mercury at rates proportional to total leaf area.We document substantial mercury accumulation in soil, biomass and resident songbirds in some of the most protected and biodiversity-rich regions of the Amazon, raising important questions about how mercury pollution constrains modern and future conservation efforts in these tropical ecosystems question.
A growing challenge for tropical forest ecosystems is artisanal and small-scale gold mining (ASGM).This form of gold mining occurs in more than 70 countries, often informally or illegally, and accounts for about 20% of the world’s gold production1.While ASGM is an important livelihood for local communities, it results in widespread deforestation2,3, extensive conversion of forests to ponds4, high sediment content in nearby rivers5,6, and is a major contributor to the global atmosphere Release of mercury (Hg) emissions and largest sources of freshwater mercury 7. Many intensified ASGM sites are located in global biodiversity hotspots, resulting in loss of diversity8, loss of sensitive species9 and human10,11,12 and apex predators13, 14 high exposure to mercury.An estimated 675–1000 tons of Hg yr-1 are volatilized and released into the global atmosphere from ASGM operations annually7.The use of large quantities of mercury by artisanal and small-scale gold mining has shifted major sources of atmospheric mercury emissions from the global north to the global south, with implications for mercury fate, transport and exposure patterns.However, little is known about the fate of these atmospheric mercury emissions and their deposition and accumulation patterns in ASGM-influenced landscapes.
The International Minamata Convention on Mercury entered into force in 2017, and Article 7 specifically addresses mercury emissions from artisanal and small-scale gold mining.In ASGM, liquid elemental mercury is added to sediments or ore to separate gold.The amalgam is then heated, concentrating the gold and releasing gaseous elemental mercury (GEM; Hg0) into the atmosphere.This is despite efforts by groups such as the United Nations Environment Programme (UNEP) Global Mercury Partnership, the United Nations Industrial Development Organization (UNIDO) and NGOs to encourage miners to reduce mercury emissions.As of this writing in 2021, 132 countries, including Peru, have signed the Minamata Convention and have begun developing national action plans to specifically address ASGM-related mercury emissions reductions.Academics have called for these national action plans to be inclusive, sustainable and holistic, taking into account socioeconomic drivers and environmental hazards15,16,17,18.Current plans to address the consequences of mercury in the environment focus on mercury risks associated with artisanal and small-scale gold mining near aquatic ecosystems, involving miners and people living near amalgam burning, and communities that consume large amounts of predatory fish .Occupational mercury exposure through inhalation of mercury vapor from the combustion of amalgam, dietary mercury exposure through consumption of fish, and mercury bioaccumulation in aquatic food webs have been the focus of most ASGM-related scientific research, including in the Amazon. Earlier studies (eg, see Lodenius and Malm19).
Terrestrial ecosystems are also at risk of mercury exposure from ASGM.Atmospheric Hg released from ASGM as GEM can return to the terrestrial landscape through three main routes20 (Fig. 1): GEM can be adsorbed to particles in the atmosphere, which are then intercepted by surfaces; GEM can be directly absorbed by plants and incorporated into their tissues; finally , GEM can be oxidized to Hg(II) species, which can be dry deposited, adsorbed into atmospheric particles, or entrained in rainwater.These pathways supply mercury to soil through fallwater (i.e., precipitation across the tree canopy), litter, and rainfall, respectively.Wet deposition can be determined by mercury fluxes in sediment collected in open spaces.Dry deposition can be determined as the sum of the mercury flux in litter and the mercury flux in fall minus the mercury flux in precipitation.A number of studies have documented mercury enrichment in terrestrial and aquatic ecosystems in close proximity to ASGM activity (see, for example, summary table in Gerson et al. 22), likely as a result of both sedimentary mercury input and direct mercury release.However, while the enhanced mercury deposition near the ASGM may be due to the burning of mercury-gold amalgam, it is unclear how this Hg is transported in the regional landscape and the relative importance of different depositional pathways near the ASGM.
Mercury emitted as gaseous elemental mercury (GEM; Hg0) can be deposited into the landscape through three atmospheric pathways.First, GEM can be oxidized to ionic Hg (Hg2+), which can be entrained in water droplets and deposited on leaf surfaces as wet or dry deposits.Second, GEMs can adsorb atmospheric particulate matter (Hgp), which is intercepted by foliage and washed into the landscape through waterfalls along with the intercepted ionic Hg.Third, GEM can be absorbed into leaf tissue, while Hg is deposited in the landscape as litter.Together with falling water and litter is considered an estimate of total mercury deposition.Although GEM may also diffuse and adsorb directly to soil and litter77, this may not be the primary route for mercury entry into terrestrial ecosystems.
We expect gaseous elemental mercury concentrations to decrease with distance from mercury emission sources.Since two of the three pathways of mercury deposition into landscapes (through fall and litter) depend on mercury interactions with plant surfaces, we can also predict the rate at which mercury is deposited into ecosystems and how severe it is for animals The risk of impact is determined by vegetation structure, as shown by observations in boreal and temperate forests in northern latitudes23.However, we also recognize that ASGM activity frequently occurs in the tropics, where canopy structure and relative abundance of exposed leaf area vary widely.The relative importance of mercury deposition pathways in these ecosystems has not been clearly quantified, especially for forests close to mercury emission sources, the intensity of which is rarely observed in boreal forests.Therefore, in this study, we ask the following questions: (1) How do gaseous elemental mercury concentrations and deposition pathways vary with the proximity of ASGM and the leaf area index of the regional canopy?(2) Are soil mercury storage related to atmospheric inputs?(3) Is there evidence of elevated mercury bioaccumulation in forest-dwelling songbirds near ASGM?This study is the first to examine mercury deposition inputs near ASGM activity and how canopy cover correlates with these patterns, and the first to measure methylmercury (MeHg) concentrations in the Peruvian Amazon landscape.We measured GEM in the atmosphere, and total precipitation, penetration, total mercury and methylmercury in leaves, litter, and soil in forest and deforested habitats along a 200-kilometer stretch of the Madre de Dios River in southeastern Peru .We hypothesized that proximity to ASGM and mining towns burning Hg-gold amalgam would be the most important factors driving atmospheric Hg concentrations (GEM) and wet Hg deposition (high precipitation).Since dry mercury deposition (penetration + litter) is related to tree canopy structure,21,24 we also expect forested areas to have higher mercury inputs than adjacent deforested areas, which, given the high leaf area index and mercury capture potential, One point is particularly worrying.Intact Amazon Forest.We further hypothesized that fauna living in forests near mining towns had higher mercury levels than fauna living far from mining areas.
Our investigations took place in the province of Madre de Dios in the southeastern Peruvian Amazon, where more than 100,000 hectares of forest have been deforested to form alluvial ASGM3 adjacent to, and sometimes within, protected lands and national reserves.Artisanal and small-scale gold mining along rivers in this western Amazon region has increased dramatically over the past decade25 and is expected to increase with high gold prices and increased connectivity to urban centers via transoceanic highways Activities will continue 3.We selected two sites without any mining (Boca Manu and Chilive, approximately 100 and 50 km from ASGM, respectively) – hereafter referred to as “remote sites” – and three sites within the mining area – hereafter referred to as “remote sites” mining site” (Fig. 2A).Two of the mining sites are located in secondary forest near the towns of Boca Colorado and La Bellinto, and one mining site is located in an intact old-growth forest on the Los Amigos Conservation Concession.Note that at the Boca Colorado and Laberinto mines of the mine, mercury vapour released from the combustion of mercury-gold amalgam occurs frequently, but the exact location and amount are unknown as these activities are often informal and clandestine; we will combine mining and mercury Alloy combustion is collectively referred to as “ASGM activity”.At each site, we installed sediment samplers in both the dry and rainy seasons in clearings (deforestation areas completely devoid of woody plants) and under tree canopies (forest areas) for a total of three seasonal events (each lasting 1- 2 months) ) Wet deposition and penetration drop were collected separately, and passive air samplers were deployed in the open space to collect GEM.The following year, based on the high deposition rates measured in the first year, we installed collectors on six additional forest plots in Los Amigos.
The maps of the five sampling points are shown as yellow circles.Two sites (Boca Manu, Chilive) are located in areas far from artisanal gold mining, and three sites (Los Amigos, Boca Colorado and Laberinto) are located in areas affected by mining, with mining towns shown as blue triangles.The illustration shows a typical remote forested and deforested area affected by mining.In all figures, the dashed line represents the dividing line between the two remote sites (left) and the three mining-affected sites (right).B Gaseous elemental mercury (GEM) concentrations at each site in the 2018 dry season (n = 1 independent sample per site; square symbols) and wet season (n = 2 independent samples; square symbols) seasons.C Total mercury concentrations in precipitation collected in forest (green boxplot) and deforestation (brown boxplot) areas during the dry season of 2018.For all boxplots, lines represent medians, boxes show Q1 and Q3, whiskers represent 1.5 times the interquartile range (n = 5 independent samples per forest site, n = 4 independent samples per deforestation site sample).D Total mercury concentrations in leaves collected from the canopy of Ficus insipida and Inga feuillei during the dry season in 2018 (left axis; dark green square and light green triangle symbols, respectively) and from bulk litter on the ground (right axis; olive green circle symbols) .Values are shown as mean and standard deviation (n = 3 independent samples per site for live leaves, n = 1 independent sample for litter).E Total mercury concentrations in topsoil (top 0-5 cm) collected in forest (green boxplot) and deforestation (brown boxplot) areas during the dry season of 2018 (n = 3 independent samples per site).Data for other seasons are shown in Figure 1.S1 and S2.
Atmospheric mercury concentrations (GEM) were in line with our predictions, with high values around ASGM activity—especially around towns burning Hg-gold amalgam—and low values in areas far from active mining areas (Fig. 2B).In remote areas, GEM concentrations are below the global average background concentration in the southern hemisphere of about 1 ng m-326.In contrast, GEM concentrations in all three mines were 2-14 times higher than in remote mines, and concentrations in nearby mines (up to 10.9 ng m-3) were comparable to those in urban and urban areas, and sometimes exceeded those in the U.S., Industrial Zones in China and Korea 27.This GEM pattern in Madre de Dios is consistent with mercury-gold amalgam burning as the main source of elevated atmospheric mercury in this remote Amazon region.
While GEM concentrations in clearings tracked proximity to mining, total mercury concentrations in penetrating waterfalls depended on proximity to mining and forest canopy structure.This model suggests that GEM concentrations alone do not predict where high mercury will be deposited in the landscape.We measured the highest mercury concentrations in intact mature forests within the mining area (Fig. 2C).Los Amigos Conservation Conservation had the highest average concentrations of total mercury in the dry season (range: 18-61 ng L-1) reported in the literature and was comparable to levels measured at sites contaminated by cinnabar mining and industrial coal combustion. Difference, 28 in Guizhou, China.To our knowledge, these values represent the maximum annual throughput mercury fluxes calculated using the dry and wet season mercury concentrations and precipitation rates (71 µg m-2 yr-1; Supplementary Table 1).The other two mining sites did not have elevated levels of total mercury compared to the remote sites (range: 8-31 ng L-1; 22-34 µg m-2 yr-1).With the exception of Hg, only aluminum and manganese had elevated throughputs in the mining area, likely due to mining-related land clearing; all other measured major and trace elements did not vary between mining and remote areas (Supplementary Data File 1 ), a finding consistent with leaf mercury dynamics 29 and ASGM amalgam combustion, rather than airborne dust, as the main source of mercury in the penetrating fall.
In addition to serving as adsorbents for particulate and gaseous mercury, plant leaves can directly absorb and integrate GEM into tissues30,31.In fact, at sites close to ASGM activity, trash is a major source of mercury deposition.Mean concentrations of Hg (0.080–0.22 µg g−1) measured in living canopy leaves from all three mining sites exceeded the published values for temperate, boreal, and alpine forests in North America, Europe, and Asia, as well as other Amazonian forests in South America, located in South America. Remote areas and near point sources 32, 33, 34.Concentrations are comparable to those reported for foliar mercury in subtropical mixed forests in China and Atlantic forests in Brazil (Fig. 2D)32,33,34.Following the GEM model, the highest total mercury concentrations in bulk litter and canopy leaves were measured in secondary forests within the mining area.However, the estimated waste mercury fluxes were highest in intact primary forest at the Los Amigos mine, likely due to the greater waste mass.We multiplied the previously reported Peruvian Amazon 35 by the Hg measured in the litter (average between wet and dry seasons) (Fig. 3A).This input suggests that proximity to mining areas and tree canopy cover are significant contributors to the mercury loads in ASGM in this region.
Data are shown in A forest and B deforestation area.The deforested areas of Los Amigos are field station clearings that make up a small portion of the total land.Fluxes are shown with arrows and expressed as µg m-2 yr-1.For the top 0-5 cm of soil, the pools are shown as circles and expressed in μg m-2.Percentage represents the percentage of mercury present in the pool or flux in the form of methylmercury.Average concentrations between dry seasons (2018 and 2019) and rainy seasons (2018) for total mercury through rainfall, bulk precipitation, and litter, for scale-up estimates of mercury loads.Methylmercury data is based on the 2018 dry season, the only year for which it was measured.See “Methods” for information on pooling and flux calculations.C Relationship between total mercury concentration and leaf area index in eight plots of Los Amigos Conservation Conservation, based on ordinary least squares regression.D Relationship between total mercury concentration in precipitation and total surface soil mercury concentration for all five sites in forest (green circles) and deforestation (brown triangles) regions, according to ordinary least squares regression (error bars show standard deviation).
Using long-term precipitation and litter data, we were able to scale measurements of penetration and litter mercury content from the three campaigns to provide an estimate of the annual atmospheric mercury flux for the Los Amigos Conservation Concession (penetration + litter amount + precipitation) for a preliminary estimate.We found that atmospheric mercury fluxes in forest reserves adjacent to ASGM activity were more than 15 times higher than in surrounding deforested areas (137 versus 9 µg Hg m-2 yr-1; Figure 3 A,B).This preliminary estimate of mercury levels in Los Amigos exceeds previously reported mercury fluxes near point sources of mercury in forests in North America and Europe (eg, coal burning), and is comparable to values in industrial China 21,36 .All told, approximately 94% of total mercury deposition in the protected forests of Los Amigos is produced by dry deposition (penetration + litter – precipitation mercury), a contribution much higher than that of most other forest landscapes worldwide.These results highlight elevated levels of mercury entering forests by dry deposition from ASGM and the importance of the forest canopy in removing ASGM-derived mercury from the atmosphere.We anticipate that the highly enriched Hg deposition pattern observed in forested areas near ASGM activity is not unique to Peru.
In contrast, deforested areas in mining areas have lower mercury levels, mainly through heavy precipitation, with little mercury input through fall and litter.Concentrations of total mercury in bulk sediments in the mine area were comparable to those measured in remote areas (Fig. 2C).Mean concentrations (range: 1.5–9.1 ng L-1) of total mercury in dry season bulk precipitation were lower than previously reported values in the Adirondacks of New York37 and were generally lower than those in remote Amazonian regions38.Therefore, the bulk precipitation input of Hg was lower (8.6-21.5 µg Hg m-2 yr-1) in the adjacent deforested area compared to the GEM, through-drop and litter concentration patterns of the mining site, and Does not reflect proximity to mining.Because ASGM requires deforestation,2,3 cleared areas where mining activities are concentrated have lower mercury inputs from atmospheric deposition than nearby forested areas, although non-atmospheric direct releases of ASGM (such as elemental mercury spills or tailings) are likely to be very high. High 22.
Changes in mercury fluxes observed in the Peruvian Amazon are driven by large differences within and between sites during the dry season (forest and deforestation) (Fig. 2).In contrast, we saw minimal intra-site and inter-site differences as well as low Hg fluxes during the rainy season (Supplementary Fig. 1).This seasonal difference (Fig. 2B) may be due to the higher intensity of mining and dust production in the dry season.Increased deforestation and reduced precipitation during dry seasons may increase dust production, thereby increasing the amount of atmospheric particles that absorb mercury.Mercury and dust production during the dry season may contribute to mercury flux patterns within deforestation compared to the forested areas of the Los Amigos Conservation Concession.
As mercury inputs from ASGM in the Peruvian Amazon are deposited into terrestrial ecosystems primarily through interactions with the forest canopy, we tested whether higher tree canopy density (ie, leaf area index) would lead to higher mercury inputs.In the intact forest of Los Amigos Conservation Concession, we collected drop drop from 7 forest plots with different canopy densities.We found that leaf area index was a strong predictor of total mercury input through fall, and the mean total mercury concentration through fall increased with leaf area index (Fig. 3C).Many other variables also affect mercury input through drop, including leaf age34, leaf roughness, stomatal density, wind speed39, turbulence, temperature, and pre-dry periods.
Consistent with the highest mercury deposition rates, the topsoil (0-5 cm) of the Los Amigos forest site had the highest total mercury concentration (140 ng g-1 in the 2018 dry season; Fig. 2E).Furthermore, mercury concentrations were enriched across the entire measured vertical soil profile (range 138–155 ng g-1 at a depth of 45 cm; Supplementary Fig. 3).The only site that exhibited high surface soil mercury concentrations during the 2018 dry season was a deforestation site near a mining town (Boca Colorado).At this site, we hypothesized that the extremely high concentrations may be due to localized contamination of elemental mercury during fusion, as concentrations did not rise at depth (>5 cm).The fraction of atmospheric mercury deposition lost to escaping from soil (i.e. mercury released into the atmosphere) due to canopy cover may also be much lower in forested areas than in deforestationed areas40, suggesting that a significant proportion of mercury is deposited to conservation. The area remains in the soil.Soil total mercury pools in the primary forest of the Los Amigos Conservation Conservation were 9100 μg Hg m-2 within the first 5 cm and over 80,000 μg Hg m-2 within the first 45 cm.
Since leaves primarily absorb atmospheric mercury, rather than soil mercury,30,31 and then transport this mercury into soil by falling, it is possible that the high deposition rate of mercury drives the patterns observed in soil.We found a strong correlation between mean total mercury concentrations in topsoil and total mercury concentrations in all forest areas, whereas there was no relationship between topsoil mercury and total mercury concentrations in heavy precipitation in deforested areas (Fig. 3D).Similar patterns were also evident in the relationship between topsoil mercury pools and total mercury fluxes in forested areas, but not in deforestation areas (topsoil mercury pools and total precipitation total mercury fluxes).
Almost all studies of terrestrial mercury pollution associated with ASGM have been limited to measurements of total mercury, but methylmercury concentrations determine mercury bioavailability and subsequent nutrient accumulation and exposure.In terrestrial ecosystems, mercury is methylated by microorganisms under anoxic conditions41,42, so it is generally believed that highland soils have lower concentrations of methylmercury.However, for the first time, we have recorded measurable concentrations of MeHg in Amazonian soils near ASGMs, suggesting that elevated MeHg concentrations extend beyond aquatic ecosystems and into terrestrial environments within these ASGM-affected areas, including those that are submerged during the rainy season. Soil and those that stay dry year round.The highest concentrations of methylmercury in the topsoil during the 2018 dry season occurred in two forested areas of the mine (Boca Colorado and Los Amigos Reserve; 1.4 ng MeHg g−1, 1.4% Hg as MeHg and 1.1 ng MeHg g−1, respectively, at 0.79% Hg (as MeHg). Since these percentages of mercury in the form of methylmercury are comparable to other terrestrial locations worldwide (Supplementary Fig. 4), the high concentrations of methylmercury appear to be due to high total Mercury input and high storage of total mercury in soil, rather than net conversion of available inorganic mercury to methylmercury (Supplementary Fig. 5). Our results represent the first measurements of methylmercury in soils near ASGM in the Peruvian Amazon. According to Other studies have reported higher methylmercury production in flooded and arid landscapes43,44 and we expect higher methylmercury concentrations in nearby forest seasonal and permanent wetlands that experience similar mercury loads. Although methylmercury Whether there is a toxicity risk to terrestrial wildlife near gold mining activities remains to be determined, but these forests close to ASGM activities may be hotspots for mercury bioaccumulation in terrestrial food webs.
The most important and novel implication of our work is to document the transport of large quantities of mercury into forests adjacent to ASGM.Our data suggest that this mercury is available in, and moves through, terrestrial food webs.In addition, significant amounts of mercury are stored in biomass and soils and are likely to be released with land-use change4 and forest fires45,46.The southeastern Peruvian Amazon is one of the most biologically diverse ecosystems of vertebrate and insect taxa on Earth.High structural complexity within intact ancient tropical forests promotes bird biodiversity48 and provides niches for a wide range of forest-dwelling species49.As a result, more than 50% of the Madre de Dios area is designated as protected land or a national reserve50.International pressure to control illegal ASGM activity in the Tambopata National Reserve has grown significantly over the past decade, leading to a major enforcement action (Operación Mercurio) by the Peruvian government in 2019.However, our findings suggest that the complexity of the forests that underlie Amazonian biodiversity makes the region highly vulnerable to mercury loading and storage in landscapes with increased ASGM-related mercury emissions, leading to global mercury fluxes through water. The highest reported measurement of the amount is based on our preliminary estimates of elevated litter mercury fluxes in intact forests near ASGM.While our investigations took place in protected forests, the pattern of elevated mercury input and retention would apply to any old-growth primary forest near ASGM activity, including buffer zones, so these results are consistent with protected and unprotected forests. Protected forests are similar.Therefore, the risks of ASGM to mercury landscapes are not only related to the direct import of mercury through atmospheric emissions, spills, and tailings, but also to the landscape’s ability to capture, store, and convert mercury into more bioavailable forms. related to potential.methylmercury, showing differential effects on global mercury pools and terrestrial wildlife depending on forest cover near mining.
By sequestering atmospheric mercury, intact forests near artisanal and small-scale gold mining can reduce mercury risks to nearby aquatic ecosystems and global atmospheric mercury reservoirs.If these forests are cleared for expanded mining or agricultural activities, residual mercury can be transferred from land to aquatic ecosystems through forest fires, escape and/or runoff45, 46, 51, 52, 53.In the Peruvian Amazon, about 180 tons of mercury are used annually in ASGM54, of which about a quarter is emitted into the atmosphere55, given the Conservation Concession at Los Amigos.This area is approximately 7.5 times the total area of protected land and nature reserves in the Madre de Dios region (about 4 million hectares), which has the largest proportion of protected land in any other Peruvian province, and these large areas of intact forest land. Partially outside the deposition radius of ASGM and mercury.Thus, mercury sequestration in intact forests is not sufficient to prevent ASGM-derived mercury from entering regional and global atmospheric mercury pools, suggesting the importance of reducing ASGM mercury emissions.The fate of large quantities of mercury stored in terrestrial systems is largely influenced by conservation policies.Future decisions on how to manage intact forests, especially in areas near ASGM activity, thus have implications for mercury mobilization and bioavailability now and in the coming decades.
Even if forests could sequester all mercury released in tropical forests, it would not be a panacea for mercury pollution, as terrestrial food webs may also be vulnerable to mercury.We know very little about mercury concentrations in biota within these intact forests, but these first measurements of terrestrial mercury deposits and soil methylmercury suggest that high levels of mercury in soil and high methylmercury may increase exposure to those living in these forests. Risks for high nutritional-grade consumers. Data from previous studies on terrestrial mercury bioaccumulation in temperate forests have found that blood mercury concentrations in birds correlate with mercury concentrations in sediments, and songbirds eating foods derived entirely from land may exhibit mercury concentrations Raised 56,57.Elevated mercury exposure in songbirds is associated with reduced reproductive performance and success, reduced offspring survival, impaired development, behavioral changes, physiological stress, and mortality58,59.If this model holds true for the Peruvian Amazon, the high mercury fluxes that occur in intact forests could lead to high mercury concentrations in birds and other biota, with possible adverse effects.This is especially concerning because the region is a global biodiversity hotspot60.These results underscore the importance of preventing artisanal and small-scale gold mining from taking place within national protected areas and the buffer zones surrounding them.Formalizing ASGM activities15,16 may be a mechanism to ensure that protected lands are not exploited.
To assess whether mercury deposited in these forested areas is entering the terrestrial food web, we measured the tail feathers of several resident songbirds from the Los Amigos Reserve (affected by mining) and the Cocha Cashu Biological Station (unaffected old birds). total mercury concentration.growth forest), 140 km from our most upstream Bokamanu sampling site.For all three species where multiple individuals were sampled at each site, Hg was elevated in birds of Los Amigos compared with Cocha Cashu (Fig. 4).This pattern persisted regardless of feeding habits, as our sample included the understory anti-eater Myrmotherula axillaris, the ant-followed anti-eater Phlegopsis nigromaculata, and the fruit-eater Pipra fasciicauda (1.8 [n = 10] vs. 0.9 μg g− 1 [n = 2], 4.1 [n = 10] vs. 1.4 μg g-1 [n = 2], 0.3 [n = 46] vs. 0.1 μg g-1 [n = 2]).Of the 10 Phlegopsis nigromaculata individuals sampled at Los Amigos, 3 exceeded EC10 (effective concentration for a 10% reduction in reproductive success), 3 exceeded EC20, 1 exceeded EC30 (see EC criteria in Evers58), and no individual Cocha Any species of Cashu exceeds EC10.These preliminary findings, with average mercury concentrations 2-3 times higher in songbirds from protected forests adjacent to ASGM activity, and individual mercury concentrations up to 12 times higher, raise concerns that mercury contamination from ASGM may enter terrestrial food webs. degree of considerable concern.These results underscore the importance of preventing ASGM activity in national parks and their surrounding buffer zones.
Data were collected at Los Amigos Conservation Concessions (n = 10 for Myrmotherula axillaris [understory invertivore] and Phlegopsi nigromaculata [ant-following invertivore], n = 46 for Pipra fasciicauda [frugivore]; red triangle symbol) and remote locations in Cocha Kashu Biological Station (n = 2 per species; green circle symbols).Effective concentrations (ECs) are shown to reduce reproductive success by 10%, 20% and 30% (see Evers58).Bird photos modified from Schulenberg65.
Since 2012, the extent of ASGM in the Peruvian Amazon has increased by more than 40% in protected areas and 2,25 or more in unprotected areas.Continued use of mercury in artisanal and small-scale gold mining can have devastating effects on the wildlife that inhabit these forests.Even if miners stop using mercury immediately, the effects of this contaminant in soils can last for centuries, with the potential to increase losses from deforestation and forest fires61,62.Thus, mercury pollution from ASGM may have long-lasting effects on the biota of intact forests adjacent to ASGM, current risks and future risks through mercury releases in old-growth forests with the highest conservation value. and reactivation to maximize the contamination potential.Our finding that terrestrial biota may be at considerable risk of mercury contamination from ASGM should provide further impetus for continued efforts to reduce mercury releases from ASGM.These efforts include a variety of approaches, from relatively simple mercury capture distillation systems to more challenging economic and social investments that will formalize the activity and reduce the economic incentives for illegal ASGM.
We have five stations within 200 km of the Madre de Dios River.We selected sampling sites based on their proximity to intensive ASGM activity, approximately 50 km between each sampling site, accessible via the Madre de Dios River (Fig. 2A).We have selected two sites without any mining (Boca Manu and Chilive, approximately 100 and 50 km from ASGM, respectively), hereafter referred to as “remote sites”.We selected three sites within the mining area, hereinafter referred to as “Mining Sites”, two mining sites in secondary forest near the towns of Boca Colorado and Laberinto, and one mining site in intact primary forest.Los Amigos Protection Concessions.Please note that at the Boca Colorado and Laberinto sites in this mining area, mercury vapour released from the combustion of mercury-gold amalgam is a frequent occurrence, but the exact location and amount are unknown as these activities are often illegal and clandestine; we will combine mining and mercury Alloy combustion is collectively referred to as “ASGM activity”.During the 2018 dry season (July and August 2018) and the 2018 rainy season (December 2018) in clearings (deforestation areas completely free of woody plants) and under tree canopies (forest areas), we Sediment samplers were installed at five sites and in January 2019) to collect wet deposition (n = 3) and penetration drop (n = 4), respectively.Precipitation samples were collected during four weeks in the dry season and two to three weeks in the rainy season.During the second year of dry season sampling (July and August 2019), we installed collectors (n = 4) in six additional forest plots in Los Amigos for five weeks, based on the high deposition rates measured in the first year, There are a total of 7 forest plots and 1 deforestation plot for Los Amigos.The distance between plots was 0.1 to 2.5 km.We collected one GPS waypoint per plot using a handheld Garmin GPS.
We deployed passive air samplers for mercury at each of our five locations during the 2018 dry season (July-August 2018) and the 2018 rainy season (December 2018-January 2019) for two months (PAS).One PAS sampler was deployed per site during the dry season and two PAS samplers were deployed during the rainy season.PAS (developed by McLagan et al. 63) collects gaseous elemental mercury (GEM) by passive diffusion and adsorption onto a sulfur-impregnated carbon sorbent (HGR-AC) via a Radiello© diffusion barrier.The diffusion barrier of PAS acts as a barrier against the passage of gaseous organic mercury species; therefore, only GEM is adsorbed to carbon 64.We used plastic cable ties to attach the PAS to a post about 1 m above the ground.All samplers were sealed with parafilm or stored in resealable double-layer plastic bags before and after deployment.We collected field blank and travel blank PAS to assess contamination introduced during sampling, field storage, laboratory storage, and sample transport.
During the deployment of all five sampling sites, we placed three precipitation collectors for mercury analysis and two collectors for other chemical analyses, and four pass-through collectors for mercury analysis at the deforestation site. collector, and two collectors for other chemical analyses.The collectors are one meter apart from each other.Note that while we have a consistent number of collectors installed at each site, during some collection periods we have smaller sample sizes due to site flooding, human interference with collectors, and connection failures between tubing and collection bottles.At each forest and deforestation site, one collector for mercury analysis contained a 500-mL bottle, while the other contained a 250-mL bottle; all other collectors for chemical analysis contained a 250-mL bottle.These samples were kept refrigerated until freezer-free, then shipped to the United States on ice, and then kept frozen until analysis.The collector for mercury analysis consists of a glass funnel passed through a new styrene-ethylene-butadiene-styrene block polymer (C-Flex) tube with a new polyethylene terephthalate Ester copolyester glycol (PETG) bottle with a loop that acts as a vapor lock.At deployment, all 250 mL PETG bottles were acidified with 1 mL trace metal grade hydrochloric acid (HCl) and all 500 mL PETG bottles were acidified with 2 mL trace metal grade HCl.The collector for other chemical analyses consists of a plastic funnel connected to a polyethylene bottle via new C-Flex tubing with a loop that acts as a vapor lock.All glass funnels, plastic funnels and polyethylene bottles were acid washed prior to deployment.We collected samples using the clean hands-dirty hands protocol (EPA Method 1669), kept samples as cold as possible until return to the United States, and then stored samples at 4°C until analysis.Previous studies using this method have shown 90-110% recoveries for laboratory blanks below the detection limit and standard spikes37.
At each of the five sites, we collected leaves as canopy leaves, grabbed leaf samples, fresh litter, and bulk litter using the clean-hands-dirty-hands protocol (EPA Method 1669).All samples were collected under a collection license from SERFOR, Peru, and imported into the United States under a USDA import license.We collected canopy leaves from two tree species found at all sites: an emerging tree species (Ficus insipida) and a medium-sized tree (Inga feuilleei).We collected leaves from tree canopies using the Notch Big Shot slingshot during the 2018 dry season, the 2018 rainy season, and the 2019 dry season (n = 3 per species).We collected leaf grab samples (n = 1) by collecting leaves from each plot from branches less than 2 m above the ground during the 2018 dry season, the 2018 rainy season, and the 2019 dry season.In 2019, we also collected leaf grab samples (n = 1) from 6 additional forest plots in Los Amigos.We collected fresh litter (“bulk litter”) in plastic mesh-lined baskets (n = 5) during the 2018 rainy season at all five forest sites and during the 2019 dry season at the Los Amigos plot (n = 5).Note that while we installed a consistent number of baskets at each site, during some collection periods, our sample size was smaller due to site flooding and human interference with the collectors.All trash baskets are placed within one meter of the water collector.We collected bulk litter as ground litter samples during the 2018 dry season, the 2018 rainy season, and the 2019 dry season.During the dry season of 2019, we also collected a large amount of litter across all of our Los Amigos plots.We refrigerated all leaf samples until they could be frozen using a freezer, then shipped to the US on ice, and then stored frozen until processing.
We collected soil samples in triplicate (n = 3) from all five sites (open and canopy) and the Los Amigos plot during the 2019 dry season during all three seasonal events.All soil samples were collected within one meter of the precipitation collector.We collected soil samples as topsoil under the litter layer (0–5 cm) using a soil sampler.Additionally, during the 2018 dry season, we collected soil cores up to 45 cm deep and divided them into five depth segments.At Laberinto, we could only collect one soil profile because the water table is close to the soil surface.We collected all samples using the clean hand-dirty hand protocol (EPA Method 1669).We refrigerated all soil samples until they could be frozen using a freezer, then shipped on ice to the United States, and then stored frozen until processing.
Use fog nests set at dawn and dusk to catch birds during the coolest times of the day.In the Los Amigos Reserve, we placed five fog nests (1.8 × 2.4) in nine locations.At the Cocha Cashu Bio Station, we placed 8 to 10 fog nests (12 x 3.2 m) in 19 locations.At both sites, we collected each bird’s first central tail feather, or if not, the next oldest feather.We store feathers in clean Ziploc bags or manila envelopes with silicone.We collected photographic records and morphometric measurements to identify species according to Schulenberg65.Both studies were supported by SERFOR and permission from the Animal Research Council (IACUC).When comparing bird feather Hg concentrations, we examined those species whose feathers were collected at the Los Amigos Conservation Concession and the Cocha Cashu Biological Station (Myrmotherula axillaris, Phlegopsis nigromaculata, Pipra fasciicauda).
To determine the Leaf Area Index (LAI), lidar data was collected using the GatorEye Unmanned Aerial Laboratory, a sensor fusion unmanned aerial system (see www.gatoreye.org for details, also available using the “2019 Peru Los Friends” June” link) 66.The lidar was collected at Los Amigos Conservation Conservation in June 2019, with an altitude of 80 m, a flight speed of 12 m/s, and a distance of 100 m between adjacent routes, so the lateral deviation coverage rate reached 75%.The density of points distributed over the vertical forest profile exceeds 200 points per square meter.The flight area overlaps with all sampling areas in Los Amigos during the 2019 dry season.
We quantified the total Hg concentration of PAS-collected GEMs by thermal desorption, fusion, and atomic absorption spectroscopy (USEPA Method 7473) using a Hydra C instrument (Teledyne, CV-AAS).We calibrated CV-AAS using the National Institute of Standards and Technology (NIST) Standard Reference Material 3133 (Hg standard solution, 10.004 mg g-1) with a detection limit of 0.5 ng Hg.We performed Continuous Calibration Verification (CCV) using NIST SRM 3133 and Quality Control Standards (QCS) using NIST 1632e (bituminous coal, 135.1 mg g-1).We divided each sample into a different boat, placed it between two thin layers of sodium carbonate (Na2CO3) powder, and covered it with a thin layer of aluminum hydroxide (Al(OH)3) powder67.We measured the total HGR-AC content of each sample to remove any inhomogeneity in the Hg distribution in the HGR-AC sorbent.Therefore, we calculated the mercury concentration for each sample based on the sum of the total mercury measured by each vessel and the entire HGR-AC sorbent content in the PAS.Given that only one PAS sample was collected from each site for concentration measurements during the 2018 dry season, method quality control and assurance was performed by grouping samples with monitoring procedure blanks, internal standards, and matrix-matched criteria.During the 2018 rainy season, we repeated the measurements of the PAS samples.Values were considered acceptable when the relative percent difference (RPD) of the CCV and matrix-matched standards measurements were both within 5% of the acceptable value, and all procedural blanks were below the limit of detection (BDL).We blank-corrected total mercury measured in PAS using concentrations determined from field and trip blanks (0.81 ± 0.18 ng g-1, n = 5).We calculated GEM concentrations using the blank-corrected total mass of adsorbed mercury divided by deployment time and sampling rate (amount of air to remove gaseous mercury per unit time; 0.135 m3 day-1)63,68, adjusted for temperature and wind from World Weather Online Average temperature and wind measurements obtained for the Madre de Dios region68.The standard error reported for the measured GEM concentrations is based on the error of an external standard run before and after the sample.
We analyzed water samples for total mercury content by oxidation with bromine chloride for at least 24 hours, followed by stannous chloride reduction and purge and trap analysis, cold vapor atomic fluorescence spectroscopy (CVAFS), and gas chromatography (GC) separation (EPA Method) 1631 of the Tekran 2600 Automatic Total Mercury Analyzer, Rev. E).We performed CCV on the 2018 dry season samples using Ultra Scientific certified aqueous mercury standards (10 μg L-1) and initial calibration verification (ICV) using NIST certified reference material 1641D (mercury in water, 1.557 mg kg-1) ) with a detection limit of 0.02 ng L-1.For the 2018 wet season and 2019 dry season samples, we used the Brooks Rand Instruments Total Mercury Standard (1.0 ng L−1) for calibration and CCV and the SPEX Centriprep Inductively Coupled Plasma Mass Spectrometry (ICP-MS) multi-element for ICV solution standard 2 A with a detection limit of 0.5 ng L-1.All standards recovered within 15% of acceptable values.Field blanks, digestion blanks and analytical blanks are all BDLs.
We freeze-dried soil and leaf samples for five days.We homogenized the samples and analyzed them for total mercury by thermal decomposition, catalytic reduction, fusion, desorption, and atomic absorption spectroscopy (EPA method 7473) on a Milestone Direct Mercury Analyzer (DMA-80).For the 2018 dry season samples, we performed DMA-80 tests using NIST 1633c (fly ash, 1005 ng g-1) and the National Research Council of Canada certified reference material MESS-3 (marine sediment, 91 ng g-1). Calibration. We used NIST 1633c for CCV and MS and MESS-3 for QCS with a detection limit of 0.2 ng Hg.For the 2018 wet season and 2019 dry season samples, we calibrated the DMA-80 using the Brooks Rand Instruments Total Mercury Standard (1.0 ng L−1).We used NIST Standard Reference Material 2709a (San Joaquin soil, 1100 ng g-1) for CCV and MS and DORM-4 (fish protein, 410 ng g-1) for QCS with a detection limit of 0.5 ng Hg.For all seasons, we analyzed all samples in duplicate and accepted values when the RPD between the two samples was within 10%.Average recoveries for all standards and matrix spikes were within 10% of acceptable values, and all blanks were BDL.All reported concentrations are dry weight.
We analyzed methylmercury in water samples from all three seasonal activities, leaf samples from the 2018 dry season, and soil samples from all three seasonal activities.We extracted water samples with trace-grade sulfuric acid for at least 24 h,69 digested leaves with 2% potassium hydroxide in methanol for at least 48 h at 55°C for at least 70 h, and digested soil by microwave with trace metal-grade HNO3 acid71,72. We analyzed the 2018 dry season samples by water ethylation using sodium tetraethylborate, purge and trap, and CVAFS on a Tekran 2500 spectrometer (EPA method 1630).We used Frontier Geosciences accredited laboratory MeHg standards and sediment QCS using ERM CC580 for calibration and CCV with a method detection limit of 0.2 ng L-1.We analyzed the 2019 dry season samples using sodium tetraethylborate for water ethylation, purge and trap, CVAFS, GC, and ICP-MS on an Agilent 770 (EPA method 1630)73.We used Brooks Rand Instruments methylmercury standards (1 ng L−1) for calibration and CCV with a method detection limit of 1 pg.All standards recovered within 15% of acceptable values for all seasons and all blanks were BDL.
At our Biodiversity Institute Toxicology Laboratory (Portland, Maine, USA), the method detection limit was 0.001 μg g-1.We calibrated DMA-80 using DOLT-5 (dogfish liver, 0.44 μg g-1), CE-464 (5.24 μg g-1), and NIST 2710a (Montana soil, 9.888 μg g-1) .We use DOLT-5 and CE-464 for CCV and QCS.Average recoveries for all standards were within 5% of acceptable values, and all blanks were BDL.All replicates were within 15% RPD.All reported feather total mercury concentrations are fresh weight (fw).
We use 0.45 μm membrane filters to filter water samples for additional chemical analysis.We analyzed water samples for anions (chloride, nitrate, sulfate) and cations (calcium, magnesium, potassium, sodium) by ion chromatography (EPA method 4110B) [USEPA, 2017a] using a Dionex ICS 2000 ion chromatograph .All standards recovered within 10% of acceptable values and all blanks were BDL.We use the Thermofisher X-Series II to analyze trace elements in water samples by inductively coupled plasma mass spectrometry.Instrument calibration standards were prepared by serial dilution of certified water standard NIST 1643f.All whitespace is BDL.
All fluxes and pools reported in the text and figures use mean concentration values for the dry and rainy seasons.See Supplementary Table 1 for estimates of pools and fluxes (average annual fluxes for both seasons) using the minimum and maximum measured concentrations during the dry and rainy seasons.We calculated forest mercury fluxes from the Los Amigos Conservation Concession as summed mercury input through drop and litter.We calculated Hg fluxes from deforestation from bulk precipitation Hg deposition.Using daily rainfall measurements from Los Amigos (collected as part of EBLA and available from ACCA on request), we calculated the average cumulative annual rainfall over the past decade (2009-2018) to be approximately 2500 mm yr-1 .Note that in the 2018 calendar year, the annual rainfall is close to this average (2468mm), while the wettest months (January, February and December) account for about half of the annual rainfall (1288mm of 2468mm) . We therefore use the average of wet and dry season concentrations in all flux and pool calculations.This also allows us to consider not only the difference in precipitation between wet and dry seasons, but also the difference in ASGM activity levels between these two seasons.Since literature values of reported annual mercury fluxes from tropical forests vary between expanding mercury concentrations from dry and rainy seasons or only from dry seasons, when comparing our calculated fluxes to literature values, we directly compare our calculated mercury fluxes , while another study took samples in both the dry and wet seasons, and re-estimated our fluxes using only dry-season mercury concentrations when another study took samples only in the dry season (e.g., 74).
To determine the annual total mercury content of throughout rainfall, bulk rainfall and litter in Los Amigos, we used the difference between the dry season (average of all Los Amigos sites in 2018 and 2019) and the rainy season (average of 2018) average total mercury concentration.For total mercury concentrations at other locations, the average concentrations between the 2018 dry season and the 2018 rainy season were used.For methylmercury loads, we used data from the dry season of 2018, the only year for which methylmercury was measured.To estimate litter mercury fluxes, we used literature estimates of litter rates and mercury concentrations collected from leaves in garbage baskets at 417 g m-2 yr-1 in the Peruvian Amazon.For the soil Hg pool in the upper 5 cm of the soil, we used the measured total soil Hg (2018 and 2019 dry seasons, 2018 rainy season) and MeHg concentrations in the 2018 dry season, with an estimated bulk density of 1.25 g cm-3 in the Brazilian Amazon75.We can only perform these budget calculations at our main study site, Los Amigos, where long-term rainfall datasets are available, and where the complete forest structure allows the use of previously collected litter estimates.
We process lidar flightlines using the GatorEye multiscale postprocessing workflow, which automatically computes clean merged point cloud and raster products, including digital elevation models (DEMs) at 0.5 × 0.5 m resolution.We used DEM and cleaned lidar point clouds (WGS-84, UTM 19S Meters) as input to the GatorEye Leaf Area Density (G-LAD) workflow, which computes calibrated leaf area estimates for each voxel (m3) ( m2) across the ground at the top of the canopy at a resolution of 1 × 1 × 1 m, and the derived LAI (sum of LAD within each 1 × 1 m vertical column).The LAI value of each plotted GPS point is then extracted.
We performed all statistical analyses using R version 3.6.1 statistical software76 and all visualizations using ggplot2.We performed statistical tests using an alpha of 0.05.The relationship between two quantitative variables was assessed using ordinary least squares regression.We performed comparisons between sites using the nonparametric Kruskal test and pairwise Wilcox test.
All data included in this manuscript can be found in the Supplementary Information and associated data files.The Conservación Amazónica (ACCA) provides precipitation data upon request.
Natural Resources Defense Council.Artisanal Gold: Opportunities for Responsible Investment – Summary.Investing in Artisanal Gold Summary v8 https://www.nrdc.org/sites/default/files/investing-artisanal-gold-summary.pdf (2016).
Asner, GP & Tupayachi, R. Accelerated loss of protected forests due to gold mining in the Peruvian Amazon.environment.reservoir.Wright.12, 9 (2017).
Espejo, JC et al.Deforestation and forest degradation from gold mining in the Peruvian Amazon: a 34-year outlook.Remote Sensing 10, 1–17 (2018).
Gerson, Jr. et al.The expansion of artificial lakes exacerbates mercury pollution from gold mining.science.Advanced.6, eabd4953 (2020).
Dethier, EN, Sartain, SL & Lutz, DA Elevated water levels and seasonal inversions of river suspended sediments in tropical biodiversity hotspots due to artisanal gold mining.Process.National Academy of Sciences.science.U.S. 116, 23936–23941 (2019).
Abe, CA et al.Modelling the effects of land cover change on sediment concentrations in the gold-mining Amazon basin.register.environment.often.19, 1801–1813 (2019).
Post time: Feb-24-2022