The Temporal Response of Trace and Toxic Elements to Landscape Indices of Doce River Basin, Brazil after the Fundão Dam Collapse

Abstract

Trace and potentially toxic elements were correlated with landscape indices in the Doce River Basin in Brazil after the Fundão dam collapse. Surface water samples were submitted to the Environmental Protection Agency (EPA3051A) methodology, and figures of merit were evaluated. The certified reference material of trace elements in water was used to check the accuracy. Statistical analyses were applied to identify the factors influencing element concentrations. The predictor variables were land use, soil types, meteorological factors, and longitudinal distances. The total concentration of arsenic, chromium, manganese, nickel, phosphorus, and lead, and dissolved aluminum, copper, and iron exceeded the maximum limits established by Brazilian Resolution, mainly in the upper Doce River during the rainy season. Rainfall was the predictor variable that most affected the response, influenced by dissolved copper and iron. The highest concentration of dissolved aluminum and total phosphorus were obtained during the dry season. Correlation analysis revealed a fair correlation (>0.47-0.67, p < 0.05) among all elements in surface water with soil type, forestry, and forest. The main sources of pollution were total manganese and dissolved aluminum and iron, originating from mining activities. The study emphasizes the impact of the dam collapse during the 2022-2023 period and stresses the importance of improving sewage treatment in the area.

Keywords:
multivariate analysis; tropical river; degree of contamination; seasonal variability; mining activities

Introduction

Rivers, essential for ecosystems and human life, are threatened by contamination from metals with high potential for toxicity and organic compounds due to anthropogenic activities like mining, sewage disposal, industrial waste, and agricultural runoff, leading to biodiversity loss, increased waterborne diseases, and higher concentrations of harmful substances.¹,²,³ In Brazil, the Samarco mine tailing disaster that occurred in November 2015 is an emblematic case of environmental contamination involving a dam failure, in which 34 million cubic meters of mining waste along 650 km into the Doce River, causing widespread environmental destruction and fatalities.⁴,⁵,⁶,⁷

After the Fundão dam collapse, several studies have investigated the effects of tailings deposition on water quality,⁸,⁹,¹⁰,¹¹,¹² aquatic and terrestrial flora and fauna,¹³,¹⁴,¹⁵,¹⁶,¹⁷ estuarine conditions,¹³,¹⁴,¹⁶ riparian soils,¹⁸,¹⁹,²⁰ vegetation,²¹,²² and others.²³,²⁴

Potentially toxic elements, as well as essential trace elements, can enter the environment through human activities such as mining and smelting. However, excessive concentrations can have harmful effects on the environment and living organisms.²⁵ Contamination of rivers by potentially toxic elements is chemically dependent on their distribution and lability in aquatic systems. These elements can be adsorbed onto suspended particles, incorporated into living organisms, or complexed with organic or inorganic compounds. They can also remain as free ions (hydrates). In all the previously described situations, toxicity can increase drastically.²¹,²⁴

Considering only studies on surface water quality, the determination of trace and potentially toxic elements in samples from the Doce River⁹,²⁶,²⁷,²⁸,²⁹,³⁰,³¹ and its tributaries²⁶,²⁸,²⁹,³⁰,³¹,³²,³³,³⁴ has been the subject of thorough investigation, with studies primarily focusing on the toxicology, contamination sources, and ecological risks associated with trace metals in the Doce River Basin (DRB), on periods immediately following the disaster²⁶,²⁷,²⁸,²⁹,³²,³³,³⁴ and more recently.⁹,¹⁶

Although there are several efforts to monitor the aquatic ecosystems of the DRB, the long-term effects of these impacts are not yet fully understood. In addition, even for the most recent works, the data from the collections were carried out in years prior to 2020.⁹,¹⁶ In this study, we monitored surface water collected from the Doce and Santo Antônio Rivers at twenty-one points sampled quarterly throughout 2022 and 2023. Pre-impact data provided by IGAM (Instituto Mineiro de Gestão das Águas, Minas Gerais Water Management Institute)³⁰ and post-impact data from a previously published dataset (2016 to 2023) were integrated for comparison with our results.

The surface water of the DRB is declared as class 2 according to the CONAMA 357/2005 Resolution,³⁵ which establishes maximum or reference values for physical, chemical and biological parameters. The maximum limits established for aluminum (Al), arsenic (As), chromium (Cr), copper (Cu), iron (Fe), manganese (Mn), nickel (Ni), phosphorus (P), and lead (Pb) are 100, 10, 50, 10, 300, 100, 25, 100 and 10 μg L⁻¹, respectively. These elements were chosen for this study over the others because their concentrations exceeded the maximum limits set by the Resolution.³⁵

Additionally, the data were explored to identify the main factors influencing element concentrations in Doce and Santo Antônio rivers, using Partial Redundancy Analysis (pRDA). This method combines multivariate regression with Principal Components Analysis (PCA), allowing the exploration of relationships between environmental variables (predictor variables) and concentrations of total and dissolved elements (response variables).³⁶ In this study, predictor variables used were land use, soil types, meteorological factors (as rainy and dry season) and longitudinal distances (5 km buffer) along the rivers. Doce River (affected by tailings) or the Santo Antônio River (not affected by tailings) were using as indicator variables for landscape indices.

The objectives of this study were to determine the total concentrations of As, Cr, Mn, Ni, P, and Pb, and dissolved Al, Cu, and Fe in surface water samples along the Doce and Santo Antônio Rivers. Furthermore, it was evaluated the spatial and seasonal trends between samples and concentrations of the monitored elements applying pRDA, as well as the correlation between elements using PCA. Finally, it was provided a brief literature review eight years after the Fundão dam collapse.

Experimental

Study area and sampling points

Located in the Southeast region of Brazil, the DRB has a drainage area of approximately 87 km² with an extension of around 890 km (86% within Minas Gerais and the remainder in the Espírito Santo). The DRB is within the Brazilian Iron Quadrangle, a region rich in rocks such as itabirites with highly concentrated ores of Fe, Mn, and Al and where mining activity is of great economic importance. The basin contains 229 municipalities with an estimated population of 3.5 million people. Iron and gold extraction is concentrated in the headwaters of the DRB. According to the Brazilian Institute of Geography and Statistics (IBGE), nearly 6,900 industries operating in the DRB in 2019 were responsible for mining and processing iron ore.³⁷,³⁸

This study was carried out in two rivers of the DRB: Doce and Santo Antônio. The former was directly affected by tailings from the collapse of the Fundão dam, and the latter was used as a reference river that had not been reached by mining tailings.

The sampling sites were designated to represent the upper and middle regions of the DRB, both in the State of Minas Gerais. Along the Doce and the Santo Antônio (reference river) Rivers, 21 sampling points were defined and divided into upstream (identified by letter A) and downstream (identified by letter B) of the tributary closest to the Doce River. Figure 1 presents a map with the location of the sampling points within the Doce (in red) and the Santo Antônio (in blue) Rivers.

Figure 1
Sampling points at the DBR along the Doce (RD) and Santo Antônio (SA) Rivers.

Sampling was conducted quarterly from April 2022 to January 2023, for a total of four collections carried out during the dry (April and July of 2022) and wet seasons (October of 2022 and January of 2023), respectively. The surface water samples were collected with a Van Dorn bottle at a depth of approximately 50 cm at three different sites of each location (independently replicates). The samples were then stored in clean polyethylene bottles of 15 mL (for dissolved analysis) and 50 mL (for total analysis), both acidified with a 0.5% v v⁻¹ nitric acid solution.³⁹

Inorganic quantification in surface water samples by inductively coupled plasma mass spectrometer (ICP-MS)

All glassware and flasks used for sample preparation were decontaminated in a 10% v v⁻¹ nitric acid (HNO₃) bath for at least 12 h. All the experiments were performed using concentrate HNO₃ and hydrochloric acid (HCl) (both from Merck, Germany) distilled in a DuoPUR distillation system (Milestone, Sorisole, Italy), and ultrapure water (resistivity higher than 18.2 MΩ cm) obtained from water purification system (Millipore Direct-Q system, Millipore, SAS -67120 Malssheim, France). Elemental quantifications were performed using an inductively coupled plasma mass spectrometer (ICP-MS) 7700 (Agilent Technologies, Tokyo, Japan). Argon (99.999%, White Martins-Praxair) was used for all the measurements. The plasma operating conditions used in ICP-MS are listed in Table S1 (Supplementary

Information (SI) section). The dissolved concentrations of Al, Cu, and Fe were determined directly in filtered samples (0.45 μm Millex-HV Syringe Filter Unit) 5-fold diluted with distilled-deionized water. For total concentrations of As, Cr, Mn, Ni, P, and Pb, approximately 9.0 mL of sample was added to the perfluoroalkoxy-alkanes (PFA) microwave tubes and microwave-acid-assisted digestion was performed in duplicate (Milestone Ethos 1 - Advanced Microwave Digestion System oven, Sorisole, Italy) using 800 μL of HNO₃ and 200 μL of HCl.³⁹,⁴⁰ Distilleddeionized water was used to prepare blank solutions for both dissolved and total samples. The microwave heating program was applied as follows: 10 min ramp to reach 170 °C and hold at the same temperature for 12 min. After cooling, the digestion tubes were opened in a laminar flow hood and the digests were diluted to 25 mL with distilled-deionized water.

The multielemental analytical calibration solutions and addition and recovery tests were prepared from adequate dilutions of monoelemental stock solutions containing 1,000 mg L⁻¹ of each analyte (Fluka, Switzerland) in 0.14 mol L⁻¹ HNO₃. The internal standard rhodium (Rh) was added at 10 μg L⁻¹ to the analytical calibration solutions, analytical blanks, and samples. The concentrations of the calibration curves and instrumental conditions for ICP-MS quantification (Table S1) are detailed in the SI section.

The following figures of merit of the method were evaluated: linearity, precision, accuracy, limits of detection, limits of quantification, and linear working range. The certified reference material of trace elements in water, NIST SRM 1640a (National Institute of Standards and Technology, Gaithersburg, MD, USA), was used to optimize the experimental procedures and evaluate the accuracy.

Statistical analysis

In this study, PCA was used to investigate the spatial and seasonal trends between samples and the concentrations of the elements studied. The data matrix consisted of 34 samples (sampling points) for each collection (a total of four) and nine variables (element concentrations). The data were processed in MATLAB software, v. 7.10 (The MathWorks, Natick, MA, USA), with the aid of PLS_Toolbox, v. 5.2.2 (Eigenvector Research, Manson, WA).

As previously described, pRDA was employed to identify the main factors influencing the concentrations of total and dissolved elements in the rivers. The geospatial analysis applied to the pRDA included a buffer of up to 5 km from each sampling point. pRDA combines multivariate regression with PCA, allowing the exploration of relationships between predictor variables and response variables.³⁶ Analogous to multivariate regression, pRDA can quantify the proportion of the variance explained by different sets of predictor variables, a method known as variance partitioning. In this study, the concentrations of Al, As, Cr, Cu, Fe, Mn, Ni, P, and Pb constituted the group of response variables, whereas the predictor variables were grouped into four sets: land use, soil types, meteorological variables (rainy and dry season), and longitudinal distance variables along the rivers, as well as the indicator variable of the Doce or Santo Antônio.

A forward selection procedure was applied for each group of predictor variables following the methodology previously described.⁴¹ The pRDA was then performed using only the significant variables identified within each group. Subsequently, the pure fractions of variation explained of each set as well as the joint explanation fraction between two or more sets were extracted. All analyses were performed using R programming language version 4.1.3. Specifically, we employed the “rda” and “varpart” functions of the “vegan” package⁴² and the “forward.sel” functions of the “adespatial” package.⁴³

MapBiomas land use and land cover (LULC)

MapBiomas provides information on land use and land cover in Brazil using remote sensing data and advanced image processing techniques. With this product, it is possible to obtain LULC maps from 1985 to the present, which are available for downloading on its virtual platform.⁴⁴ The available data are organized into different LULC, allowing detailed analyses of changes and trends over time. The classes of MapBiomas are as follows: (i) forest; (ii) non-forest natural formation; (iii) agriculture; (iv) non-vegetated area; and (v) water bodies. Notably, each class can be subdivided into more detailed classes. For example, the forest class is subdivided into forest formation, savannah formation, mangrove, flooded forest, and arboreal restinga. LULC has a spatial resolution of 30 meters and a web platform where users can access, visualize, and download data.⁴⁵ The accuracy analysis of the project is predetermined by statistical sampling techniques, consisting of approximately 75,000 independent samples by technicians trained in the visual interpretation of landsat images, showing a global accuracy that ranges from 70 to 95% across all Brazilian biomes.⁴⁴ All the geospatial analyses also included a buffer of up to 5 km from each sampling point.

Altimetry and soil data

The soil classification data used were derived from SEMAD (Secretaria Estadual de Meio Ambiente e Desenvolvimento Sustentável, State Department of Environment and Sustainable Development) and Universidade Federal de Viçosa (UFV),⁴⁶ which refers to a set of information and maps that describing the soil classification for Minas Gerais. The main feature of the SEMAD/UFV soil classification data is the taxonomic classification of soils according to the Brazilian Soil Classification System (SiBCS)⁴⁷ or USDA Soil Classification System (United States Department of Agriculture).⁴⁸ The delimitation of the influence areas and the estimation of land use for the 21 sampling points in the DRB show the geographical distribution of these points along the basin and provide an overview of the spatial coverage of the analyses carried out in a 5 km buffer.

Degree of contamination (C_d)

C_d indicates the collective detrimental impact of heavy metals (HM) on surface water and is determined as follows:

(1) ${\text{C}}_{\text{d}}=\sum \text{Cfi}$

(2) $\text{Cfi}=\left(\frac{\text{HMi}}{\text{MACi}}\right)-1$

where Cfi is the contamination factor for the i^th HM and MAC is the maximum admissible concentration of the i^th HM. The categories used to represent HM pollution based on C_d were as follows: < 1 low, 1-3 moderate and > 3 high pollution of HMs in the surface water body.⁴⁹

During the preparation of this work, the authors used Paperpal / Preflight to improve English writing. After using this tool/service, the authors reviewed and edited the content as needed and took full responsibility for the content of the publication.

Results and Discussion

Chemical characterization of dissolved and total metals and metalloids in the surface water by ICP-MS

Limits of detection (LOD) and quantification (LOQ) were calculated considering three times and ten times the standard deviation of 10 independent measurements from blank solutions, respectively. The isotopes, mode of acquisition, linear correlation coefficient, slopes of the analytical curves, and LOQ obtained for all analytes are shown in Table 1. The LOQ values obtained for dissolved Al, Cu, and Fe, and total As, Cr, Mn, Ni, P, and Pb were suitable for Brazilian Resolution.³⁵

Table 1
Analytical performance parameters for Al, As, Cr, Cu, Fe, Mn, Ni, P, and Pb in surface water samples by ICP-MS

Isotope	Acquisition mode	R2	Slope / (cps L μg−1)	LOQ / (μg L−1)	MLa / (μg L−1)
27Al	He	0.9929	0.0006	30	100
31P	no gas	0.9998	0.0018	12	100
52Cr	He	0.9992	0.1233	1.2	50
55Mn	He	0.9997	0.0476	1.1	100
56Fe	He	0.9996	0.0435	3.7	300
58Ni	He	0.9995	0.1628	3.1	25
65Cu	He	0.9996	0.0992	0.6	9
75As	He	0.9993	0.0204	0.33	10
208Pb	no gas	0.9999	0.0858	0.36	10

• ^a Maximum limits at Brazilian Resolution.³⁵ LOQ: limit of quantification; R²: determination coefficient.

NIST SRM 1640a was used to test the precision and accuracy of the method for all elements. The recoveries obtained ranged from 87 to 121% for all analytes, and the determined values were not statistically different from the reference concentrations at 95% confidence level. Table S2 (SI section) shows the results obtained. Repeatability was evaluated by analyzing six independent measurements of the sample spiked with the multi-element stock standard to the same three levels performed to addition and recovery tests for each target element (see SI section). Recoveries ranging from 80 to 125% were observed by spike experiments at three levels and repeatability was demonstrated by a precision ≤ 10% relative standard deviation (RSD) considering all samples performed using internal standardization. All recovery percentages obtained were always within the acceptable range of INMETRO (Instituto Nacional de Metrologia, Qualidade e Tecnologia -National Institute of Metrology, Quality and Technology, Brazil).⁵⁰

Figure 2 presents the concentrations determined by ICP-MS for dissolved Al, Cu, and Fe and total As, Cr, Mn, Ni, P, and Pb considering the four collections performed quarterly at 21 points divided into A (upstream) and B (downstream) of the river tributaries to the Doce River (totaling 34 sampling points, 25 in the Doce River and 9 in the Santo Antônio River). In Figure 2 it is possible to observe that especially for upper Doce and Santo Antônio (SA18, 19 and, 21) Rivers, Al, As, Cr, Cu, Fe, Mn, Ni, P, and Pb were 3.5, 2, 1.3, 8, 7, 11, 5, 3, and 5 times higher than the maximum limits established by the Brazilian Resolution,³⁵ i.e., 100, 10, 50, 10, 300, 100, 25, 100 and 10 μg L⁻¹, respectively.

Figure 2
Concentration determined for Al, As, Cr, Cu, Fe, Mn, Ni, P, and Pb in the 34 sampling points along the Doce river basin in four quarterly collections. (■) April/2022. (•) July/2022. (▲) October/2022. (▼) January/2023. The red line highlights the maximum limit established by the Brazilian Resolution CONAMA 357/2005.³⁵ RD: Doce River. SA: Santo Antonio River. (■) Upper Doce River. (■) Middle Doce River. (■) Santo Antonio River.

In general, in the upper Doce River, concentrations exceeding the maximum limits of the Brazilian Resolution³⁵ were observed for all investigated elements in the rainy season. Among the sampling points along the Doce River, higher concentrations of all analytes were observed in the upper section (RD1 to RD7) than in the middle section (RD8 to RD15). Elevated concentrations were recorded during the rainy season (October 2022 and January 2023). On average, across all sampling points, the concentrations of As-Mn, Cu, Cr-Ni, and Pb were approximately two-, three-, six-, and nine-fold higher, respectively, during the rainy season than during the dry season (April and July).

The levels of dissolved Fe and total Mn exceeded the maximum permitted concentrations of the Resolution³⁵ at most of the monitoring points, independent of the season. High concentration of total Ni, Mn, and Pb, exceeding the maximum limits, were especially found in the Santo Antônio River during the rainy season, particularly in January.

The DRB is located in the Iron Quadrangle, which is a relevant mining area in the State of Minas Gerais. Therefore, high background concentrations of Al, Fe, and Mn are usual in the total or dissolved fractions, because these metals are naturally present in the geological composition.⁵¹ During the rainy season, increased tributary flows can mobilize riverbed sediments, leading to resuspension and greater availability of these elements. Additionally, following the collapse of the Fundão dam, a significant increase in the concentrations of Al, Fe, and Mn, along with other potentially toxic elements, such As, Ni, and Pb from the mining tailings, is expected during the rainy season.

Conversely, the highest P concentration was observed during the dry season. This behavior may be associated with an increase in the concentration of sewage discharged into the river. The middle region of the DRB encompasses the largest urban centers in the basin with intense industrial activities, particularly in the steel and cellulose manufacturing industries.⁵² The watercourses of the rivers serve as receiving, transporting, and self-purifying channels for waste and effluents generated by economic activities, as well as domestic sewage from most municipalities situated within the DRB.⁵³,⁵⁴

The lack of adequate domestic sewage treatment is a major challenge in the basin (see Figure S1, SI section). Approximately 68% of the sewage generated by municipalities is discharged directly into watercourses without treatment. Only forty-four out of the 211 cities in the basin could treat at least 30% of the collected sewage. The consequences of insufficient sewage treatment are particularly evident in parts of the basin’s rivers, especially in the tributaries and middle stretch of the Doce River.⁵⁴

Principal component analysis was used to interpret elementary data. PCA was applied to the autoscaled dataset and the results with two principal components (PCs) are shown in Figure 3, explaining 62% of the total experimental data variance (PC1 47% and PC2 15%). The scores of the first principal component (Figures 3a) showed a tendency to separate the samples collected in January/23 (rainy season) with contributions of Cr, Mn, Ni, As, and Pb concentrations (Figure 3b) compared to the other sampling periods. This result reinforces the seasonal and temporal variability in water quality. The second principal component (Figures 3c and 3d) highlights, on the positive side, the concentrations of Al and Fe for the samples collected in April/22 and October/22 compared to the other collections. Moreover, on the negative side is the concentration of P, where we can see the grouping of the samples collected in July/22, dry season, with less river flow. Although the levels of these three elements were higher than the legal limits in several of the samples, they stood out in these two samplings, possibly because there were no other events such as resuspension in the rainy season and/or an increase in the concentration of sewage in the dry season.

Figure 3
PCA of the dataset for 9 elemental concentrations obtained in four collections. PC1-2 scores plots (a) and (c); PC1-2 loading plots (b) and (d).

Partial redundancy and correlation analysis

pRDA was used to evaluate the correlation between the total (As, Cr, Mn, Ni, P, and Pb) and dissolved (Al, Cu, and Fe) concentrations as a group of response variables (considering the four collections) and the four predictor variables, i.e., land use, soil type, rainy and dry season, and longitudinal distance variables (Figure 4).

Figure 4
Partial Redundancy Analysis (pRDA) in the Doce (▬) and Santo Antônio (▬) Rivers for the predictor variables: (a) longitudinal distance (5 km from riverside); (b) land use and occupation (forestry, forest, mosalic, rock outcrop) (c) soil types (PVAe22, LAd2, CXbe13, PVe1, LVd17, PVe14, LVAd38 and CYbe1) (d) set of four predictor variables (longitudinal distance, land use and occupation, soil types and meteorology (rainfall)). Soil types: PVAe22 (yellow-red argisol), LAd2 (yellow oxisol), CXbe13 (haplic cambisol), PVe1 (red argisol), LVd17 (red oxisol), PVe14 (red argisol), LVAd38 (yellow-red oxisol) and CYbe1 (fluvic cambisol).

The distance from the beginning of the river and a dichotomous variable that indicate the Doce Basin (Doce and Santo Antônio Rivers) (Figure 4a) were significant, with an adjusted R² value of 9.9%. All the elements detected, except Al and P, were negatively associated with distance, indicating that their concentrations decreased along the rivers (Figure 4a). In general, samples collected from Doce River showed higher values than those collected from Santo Antônio River, except for the total concentrations of As and Mn (Figure 4a), which were not grouped. For As, the concentrations were always lowest compared to the legal limits, except for RD1A and RD3A. On the other hand, for Mn exceeding values were registered in all RDB, independently of the season (Figure 2). For this multivariate analysis, the four collections were used for correlation analysis. Thus, it is possible to say that Al and P obtained the highest concentrations in the dry season, an opposite profile when compared to the group of variable responses (Figure 4d).

In Figure 4b the significant variables were forestry, rocky outcrop, forest, and mosaic (agriculture and pasture). Variables with a predictive capacity of at least 1% were selected and these variables together explained 11.8% of the total variance of the matrix of response variables. Together, these four variables had an adjusted R² value of 9.4%. It was possible to observe the opposite position of the pair forestry and forest variables with mosaic and rocky outcrops. The concentrations of all elements, except Al, were positively influenced by forestry and forest pairs. Forest has a positive relationship with silviculture as both involve the management of forested areas, although with different objectives. Native forests provide essential ecosystem services, such as biodiversity conservation and climate regulation, whereas silviculture is focused on commercial production, contributing to vegetation cover and environmental protection. Higher concentrations of Mn were observed in areas with a predominance of forestry, which can be associated with the use of fertilizers (Figure 4d).⁵⁵,⁵⁶ Dissolved Al showed a strong correlation with mosaic and rock outcrop areas (Figure 4b), as its bioavailability is enhanced by the natural presence of ironrich soils in the Iron Quadrangle region.

In Figure 4c, the significant predictor variables for soil types were PVAe22 (yellow-red argisol), LAd2 (yellow oxisol), CXbe13 (haplic cambisol), PVe1 (red argisol). Together, these four variables have an adjusted R² value of 12.5%. The opposition between the pair of variables PVAe22 (yellow-red argisol) and LAd2 (yellow oxisol) with the pair of CXbe13 (haplic cambisol) and PVe1 (red argisol), can be characterized as soil types of Doce River (left quadrant) and Santo Antônio River (concentrated in the right quadrants), respectively. The concentrations of all the elements quantified, except Al, were positively influenced by the first pair (PVAe22 and LAd2). From the dataset that included the four most important land uses and soil types, accumulated precipitation, longitudinal distance, and the indicator variable of the river, these variables were selected. Then, a redundancy analysis was conducted with the significant variables that explained 35.6% of the total variance in the response variable matrix (Figure 4d). Two groups of variables were observed, and the first group consists of forestry, Doce River, and soil types PVAe22 and LAd2. This group is opposite to the second one, formed by the fraction of rock outcrop cover, longitudinal distance along the river, and soil type CXbe13. Samples collected from the Doce River concentrated in the region of the first group indicate that forestry is present as the most significant land use, along with soil types PVAe22 (yellow-red argisol) and LAd2 (yellow oxisol), whereas samples collected from the Santo Antônio River are concentrated in the region of the second group (Figure 4d), where rock outcrops predominate as the most significant land use, along with soil type CXbe13 (haplic cambisol).

A singular profile was obtained for variable rainfall, as shown in Figure 4d, related to the accumulated precipitation. The concentrations of all elements detected, except Al and P, showed a positive association with the accumulated precipitation and with the variables of the first group, corresponding to the samples collected in the Doce River (red points). Higher concentrations of Al were quantified in samples associated with soil type CXbe13 (Figure 4c). Phosphorus showed a low association with all selected variables. As previously described, Al and P registered the highest concentrations in the dry season, which was opposite of all groups. The pRDA results of samplings (Figure S2) showed the same pRDA results obtained from the Doce and Santo Antônio Rivers.

As shown in Figure 5, the four collections were correlated with two predictor variables: average temperature and rainfall (accumulated precipitation) with an adjusted of R² of 21.3%. Because only the rainfall was significant, a single RDA axis was used to construct the graph. Thus, the RDA1 axis was used as the horizontal axis and PC1 was used as vertical axis, referring to the first component of the residual variation.

Figure 5
Partial Redundancy Analysis (pRDA) considering the meteorological predictor variable for (a) the samplings carried out from (▬) April/2022, (▬) July/2022, (▬) October/2022; and (▬) January/2023; and (b) in the (▬) Doce and (▬) Santo Antônio rivers.

It can be observed in Figure 5a the distribution profile of the elements according to the samplings carried out from April 2022 to January 2023. All elements, except Al and P were positively associated with the accumulated precipitation. During the dry season, the highest concentrations were observed to Al and P in the samples collected in April 2022 and January 2023. However, as the two rivers belong to the same basin and probably have similar volumes during the rainy season, it was not possible to note significant differences between them (see Figure 5b). However, the highest concentrations were observed in the upper Doce River (Figure 2).

In Figure S3 (SI section), a positive correlation is shown between the rock outcrop and PVe1 (red argisol) soil type and between land use with forest and forestry. There is a negative correlation between the longitudinal distance of the rivers and land use and forest, indicating that the points further upstream were more affected by forest. There was a positive and significant correlation between forestry and the soil type LAd2 (yellow oxisol), and between the soil type CXbe13 (haplic cambisol) and the longitudinal distance of the rivers.

Literature overview eight years after the Fundão dam collapse for surface water

Table S3 (SI section) shows the average concentration (minimum and maximum levels) determined for the nine elements investigated in comparison with data obtained by IGAM (for surface water sampling from 2010 to 2014, April/2015, and November/2015)³⁰ and by Fundação Renova (for water sampling from 2018 to 2022).³¹ Additionally, the comparison also included a revision of the literature between 2016 and 2023 (for surface water samples collected from 2015 to 2019).⁹,¹⁶,²⁶,²⁷,²⁸,²⁹,³²,³³,³⁴

In general, the elements that presented the highest concentrations immediately after the Fundão dam collapse (samplings carried out in November/2015) were As, Fe, Mn, and Pb,⁵⁷,⁵⁸,⁵⁹ with an emphasis on the marked increase in the contents of dissolved Fe³⁰,³² and Mn³⁰ exceeding 37 and 910 times, respectively, the maximum levels of the Brazilian Resolution.³⁵ In comparison to the maximum levels determined by IGAM in November 2015, the concentration values in this study for Cr, Cu, Fe, Ni, and Pb were approximately 1.7, 20, 1.8, 10, and 1.5 times higher. The Al and Mn concentrations were 1.7 and 8 times lower, respectively, probably in part to return to the region’s baseline value, which remains high.

Figure S4 (SI section) shows the boxplots obtained from these results (red points) compared to the literature.²⁹,³⁰ Despite the highest concentrations of Cr, Cu, P, and Pb, the values observed in our study were not statistically different from the values found in the literature at 95% confidence level, even under different sampling sizes proposed for each study (different collection points in different rivers from DRB). The IGAM data represented sampling at fourteen points (eleven in the Doce River, one in the Carmo River and one in the Galaxo do Norte River). The data provided by Fundação Renova, on the other hand, represent sampling at nine points (seven on the Doce River, one on the Carmo River and one on the Galaxo do Norte River). Other studies have provided sampling in the Doce River,⁹,²⁷ Carmo River,³³,³⁴ Galaxo do Norte River,³²,³³ or in the three rivers consecutively.²⁶,²⁸,²⁹

Eight sampling points in our study coincide with Fundação Renova³¹ along the Doce River. Comparing our results for samplings carried out in 2022 with those of Fundação Renova, it was observed that the concentrations of Al were similar. Conversely, the concentrations of As-Cu, Fe-Ni, and Pb stand out as having maximum concentrations equal to or higher than ten and five times, respectively. Compared with literature (Table S2), the concentrations of As, Cr, Cu, Ni, and Pb were higher than those previously reported.⁹,¹⁶,²⁸,³²

Degree of contamination was calculated for all the data sets, including our results. The values for Cr, Cu, and P were lower than one, indicating a low potential for pollution by these elements in the DRB. For As, Ni, and Pb values were around 5, indicating that these elements are related to the pollution in the DRB. The values for Al and Fe were around 30 whereas for Mn, the degree of contamination was 60.3. Values higher than 3 are associated with potential sources of pollution. This suggests that the rainy season may be the mainly variable responsible for resuspending tailings and contributing to the high C_d values, as previously described. Activities in the DRB, especially those related to mining, affect the surface water in the Iron Quadrangle region. Mitigation actions must be continuously implemented and revised according to environmental conditions. For example, alarmingly high concentrations of Mn suggest a high ecological risk for aquatic life because this metal can bioaccumulate in the livers of fish that are highly consumed by the riparian community. Manganese has been linked to dementia, and its concentration was elevated in all DRB samples, regardless of the sampling point.⁶⁰

Conclusions

The proposed method was validated by presenting adequate figures of merit enabling the determination of total and dissolved metal(loid)s in surface water samples along the Doce and Santo Antônio Rivers using ICP-MS. According to the Brazilian Resolution, concentrations that exceeded the maximum values of total and dissolved elements were observed mainly in the upper Doce River during the rainy season. The correlation between element concentrations and spatial and seasonal trends was observed using PCA. Our results demonstrated the enormous

influence of rainfall, compared to other predictor variables. This profile reinforces the effect of tailings resuspension and its impact on the riparian area of the DRB, increasing the seasonal and temporal variability of water quality. Additionally, the lack of adequate domestic sewage treatment in the basin is a challenge. Among the predictor variables along the rivers, correlation analysis showed a fair correlation (>0.47-0.67, p < 0.05) among all elements with soil type, forestry, and forest.

Eight years after the dam collapse, this work shows that mining activities affect the surface waters of the DRB. The degree of contamination reflects environmental risks associated with total Mn and dissolved Al and Fe, followed by the potentially toxic elements As, Ni, and Pb. Comparing our results with the literature, no statistically significant differences were observed at the 95% confidence level. The findings of this study emphasize the importance of continuous environmental monitoring for assessing and mitigating ongoing impacts.

Data Availability Statement

Data are available upon reasonable request, obeying informed consent policies of anonymity and confidentiality.

Supplementary Information

Supplementary information (presentation of the instrumental conditions, multivariate analysis and an overview of the literature) is available free of charge at http://jbcs.sbq.org.br as PDFfile.

Acknowledgments

The authors gratefully acknowledge the financial support provided by Fundação de Amparo à Pesquisa do Estado de Minas Gerais (FAPEMIG - APQ-00208-19). We also thank Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq - 140360/2021-2), and Coordenação de Aperfeiçoamento de Pessoal de Nível Superior (CAPES - Finance Code 001).

References

• ¹ Reis, M. M.; Santos, L. D. T.; da Silva, A. J.; de Pinho, G. P.; Montes, W. G.; Arch. Environ. Contam. Toxicol. 2019, 77, 527. [Crossref]
  » Crossref
• ² Baptista, M.; Cardoso, A.; Revista da Universidade Federal de Minas Gerais 2016, 20, 124. [Crossref]
  » Crossref
• ³ da Silva, G. J.; Borges, A. C.; Moreira, M. C.; Rosa, A. P.; J. Environ. Manage. 2022, 317, 115402. [Crossref]
  » Crossref
• ⁴ Andrade, G. F.; Paniz, F. P.; Martins, Jr. A. C.; Rocha, B. A.; Lobato, A. K. S.; Rodrigues, J. L.; Cardoso-Gustavson, P.; Masuda, H. P.; Batista, B. L.; Chemosphere 2018, 193, 892. [Crossref]
  » Crossref
• ⁵ Neves, L. P.; Legislação Brasileira em Seguranças de Barragens, https://ibram.org.br/wp-content/uploads/2021/02/e-book-livre-legislacao-federal-brasileira-em-seguranca-de-barragens-autor-luiz-paniago-neves.pdf, accessed in July 2025.
  » https://ibram.org.br/wp-content/uploads/2021/02/e-book-livre-legislacao-federal-brasileira-em-seguranca-de-barragens-autor-luiz-paniago-neves.pdf
• ⁶ Almeida, C. A.; de Oliveira, A. F.; Pacheco, A. A.; Lopes, R. P.; Neves, A. A.; de Queiroz, M. E. L. R.; Chemosphere 2018, 209, 411. [Crossref]
  » Crossref
• ⁷ Aires, U. R. V.; da Silva, D. D.; Fernandes Filho, E. I.; Rodrigues, L. N.; Uliana, E. M.; Amorim, R. S. S.; Ribeiro, C. B. M.; Campos, J. A.; J. Environ. Manage. 2022, 323, 116207. [Crossref]
  » Crossref
• ⁸ Instituto Mineiro de Gestão das Água (IGAM); Acompanhamento da Qualidade das Águas do Rio Doce Após o Rompimento da Barragem da Samarco no distrito de Bento Rodrigues -Mariana/MG, 2018, http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/4400, accessed in July 2025.
  » http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/4400
• ⁹ Yamamoto, F. Y.; Pauly, G. F. E.; Nascimento, L. S.; Fernandes, G. M.; Santos, M. P.; Figueira, R. C. L.; Cavalcante, R. M.; Grassi, M. T.; Abessa, D. M. S.; J. Hazard. Mater. Adv. 2023, 9, 100250. [Crossref]
  » Crossref
• ¹⁰ Costa, L. R. C.; Fioravante, I. F.; Cleaner Chem. Eng. 2024, 10, 100134. [Crossref]
  » Crossref
• ¹¹ Kütter, V. T.; Martins, G. S.; Brandini, N.; Cordeiro, R. C.; Almeida, J. P. A.; Marques, E. D.; J. Trace Elem. Miner. 2023, 5, 100084. [Crossref]
  » Crossref
• ¹² dos Santos, G. R.; Maia, L. C.; Lobo, F. A.; Santiago, A. F.; da Silva, G. A.; Environ. Pollut. 2023, 334, 122174. [Crossref]
  » Crossref
• ¹³ Fernandes, G. W.; Goulart, F. F.; Ranieri, B. D.; Coelho, M. S.; Dales, K.; Boesche, N.; Bustamante, M.; Carvalho, F. A.; Carvalho, D. C.; Dirzo, R.; Fernandes, S.; Galetti, P. M.; Millan, V. E. G.; Mielke, C.; Ramirez, J. L.; Neves, A.; Rogass, C.; Ribeiro, S. P.; Scariot, A.; Soares-Filho, B.; Natureza e Conservação 2016, 14, 35. [Crossref]
  » Crossref
• ¹⁴ Vilar, C. C.; Andrades, R.; Szablak, F. T.; Guabiroba, H. C.; Pichler, H. A.; Bastos, K. V. ; de Lima, L. R. S.; Bastos, P. G. P.; Martins, R. F.; Rodrigues, V. L. A.; Hostim-Silva, M.; Joyeux, J.-C.; Mar. Environ. Res. 2022, 175, 105565. [Crossref]
  » Crossref
• ¹⁵ da Silva, A. P. V.; Silva, A. O.; de Lima, F. R. D.; Benedet, L.; Franco, A. J.; de Souza, J. K.; Ribeiro, Jr. A. C.; Batista, É. R.; Inda, A. V.; Curi, N.; Guilherme, L. R. G.; Carneiro, M. A. C.; Environ. Res. 2022, 215, 114321. [Crossref]
  » Crossref
• ¹⁶ Longhini, C. M.; Rodrigues, S. K.; Costa, E. S.; da Silva, C. A.; Cagnin, R. C.; Gripp, M.; Lehrback, B. D.; Mill, G. N.; de Oliveira, E. M. C.; Hermogenes, C. C. M.; Rodrigues, D. G. F.; David, A. M.; Gramlich, K. C.; Bisi Júnior, R. C.; Gomes, A. A. P.; da Silva Filho, J. P.; Almeida, J. F.; de Souza, K. F.; Luz, Jr. W. A. R.; Poleze, L. M. B.; Barros, R. R.; Rigo, D.; Ghisolfi, R. D.; Neto, R. R.; Sá, F.; Sci. Total Environ. 2022, 803, 149883. [Crossref]
  » Crossref
• ¹⁷ Miguel, C.; Vianna, M. R. M.; de Deus Santos, M. R.; Mar. Pollut. Bull. 2023, 196, 115601. [Crossref]
  » Crossref
• ¹⁸ Ferreira, G. W. D.; Ribeiro, B. T.; Weindorf, D. C.; Teixeira, B. I.; Chakraborty, S.; Li, B.; Guilherme, L. R. G.; Scolforo, J. R. S.; Environ. Monit. Assess. 2021, 193, 203. [Crossref]
  » Crossref
• ¹⁹ Viana, L. M. S.; Pestana, I. A.; Carvalho, C. E. V; Salomão, M. S. M. B.; Arch. Environ. Contam. Toxicol. 2020, 79, 343. [Crossref]
  » Crossref
• ²⁰ Quadra, G. R.; Roland, F.; Barros, N.; Malm, O.; Lino, A. S.; Azevedo, G. M.; Thomaz, J. R.; Andrade-Vieira, L. F.; Praça-Fontes, M. M.; Almeida, R. M.; Mendonça, R. F.; Cardoso, S. J.; Guida, Y. S.; Campos, J. M. S.; Chemosphere 2019, 215, 753. [Crossref]
  » Crossref
• ²¹ Esteves, G. F.; Bressanin, L. A.; de Souza, K. R. D.; da Silva, A. B.; Mantovani, J. R.; Marques, D. M.; Magalhães, P. C.; Pasqual, M.; de Souza, T. C.; Environ. Sci. Pollut. Res. 2020, 27, 38662. [Crossref]
  » Crossref
• ²² Coelho, D. G.; Marinato, C. S.; de Matos, L. P.; de Andrade, H. M.; da Silva, V. M.; Neves, P. H. S.; de Oliveira, J. A.; Integr. Environ. Assess. Manage. 2020, 16, 596. [Crossref]
  » Crossref
• ²³ Zebral, Y. D.; Costa, P. G.; de Souza, M. M.; Bianchini, A.; Sci. Total Environ. 2022, 807, 151077. [Crossref]
  » Crossref
• ²⁴ Cavalheiro Paulelli, A. C.; Cesila, C. A.; Devóz, P. P.; Ruella de Oliveira, S.; Bianchi Ximenez, J. P.; Pedreira Filho, W. R.; Barbosa, Jr. F.; Environ. Res. 2022, 205, 112524. [Crossref]
  » Crossref
• ²⁵ Albuquerque, F. E. A.; Minervino, A. H. H.; Miranda, M.; Herrero-Latorre, C.; Barrêto, Jr. R. A.; Oliveira, F. L. C.; Sucupira, M. C. A.; Ortolani, E. L.; López-Alonso, M.; Sci. Total Environ. 2020, 732, 138983. [Crossref]
  » Crossref
• ²⁶ Cordeiro, M. C.; Garcia, G. D.; Rocha, A. M.; Tschoeke, D. A.; Campeão, M. E.; Appolinario, L. R.; Soares, A. C.; Leomil, L.; Froes, A.; Bahiense, L.; Rezende, C. E.; de Almeida, M. G.; Rangel, T. P.; de Oliveira, B. C. V.; de Almeida, D. Q. R.; Thompson, M. C.; Thompson, C. C.; Thompson, F. L.; Sci. Total Environ. 2019, 654, 1209. [Crossref]
  » Crossref
• ²⁷ de Carvalho, G. O.; Pinheiro, A. A.; de Sousa, D. M.; Padilha, J. A.; Souza, J. S.; Galvão, P. M.; Paiva, T. C.; Freire, A. S.; Santelli, R. E.; Malm, O.; Torres, J. P. M.; Orbital 2018, 10, 299. [Crossref]
  » Crossref
• ²⁸ Hatje, V.; Pedreira, R. M. A.; de Rezende, C. E.; Schettini, C. A. F.; de Souza, G. C.; Marin, D. C.; Hackspacher, P. C.; Sci. Rep. 2017, 7, 10706. [Crossref]
  » Crossref
• ²⁹ Vergilio, C. S.; Lacerda, D.; Souza, T. S.; de Oliveira, B. C. V.; Fioresi, V. S.; de Souza, V. V.; Rodrigues, G. R.; Barbosa, M. K. A. M.; Sartori, E.; Rangel, T. P.; de Almeida, D. Q. R.; de Almeida, M. G.; Thompson, F.; de Rezende, C. E.; Sci. Total Environ. 2021, 756, 143697. [Crossref]
  » Crossref
• ³⁰ Instituto Mineiro de Gestão das Água (IGAM); Séries Históricas, http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/405, accessed in July 2025.
  » http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/405
• ³¹ Fundação RENOVA; Download de Dados, https://monitoramentoriodoce.org/download-dos-dados/, accessed in July 2025.
  » https://monitoramentoriodoce.org/download-dos-dados/
• ³² Segura, F. R.; Nunes, E. A.; Paniz, F. P.; Paulelli, A. C. C.; Rodrigues, G. B.; Braga, G. Ú. L.; dos Reis Pedreira Filho, W.; Barbosa, F.; Cerchiaro, G.; Silva, F. F.; Batista, B. L.; Environ. Pollut. 2016, 218, 813. [Crossref]
  » Crossref
• ³³ Foesch, M. D. S.; Francelino, M. R.; Rocha, P. A.; Gomes, A. R. L.; Floresta Ambient. 2020, 27, e20180132. [Crossref]
  » Crossref
• ³⁴ Silva, D.; Bellato, C.; Marques Neto, J.; Fontes, M.; Quim. Nova 2018, 41, 857. [Crossref]
  » Crossref
• ³⁵ Conselho Nacional do Meio Ambiente (CONAMA); Resolução No. 357, de 17 de março de 2005, Dispõe Sobre a Classificação dos Corpos de Água e Diretrizes Ambientais para o Seu Enquadramento, bem como Estabelece as Condições e Padrões de Lançamento de Efluentes, e Dá outras Providências; Diário Oficial da União (DOU), Brasília, No. 53, de 18/03/2005, p. 58-63. [Link] accessed in March 2025
  » Link
• ³⁶ Borcard, D.; Gillet, F.; Legendre, P.; Numerical Ecology with R; Springer: New York, 2011. [Link] accessed in July 2025.
  » Link
• ³⁷ Agência Nacional de Águas e Saneamento Básico (ANA); Rio Doce, https://www.gov.br/ana/pt-br/sala-de-situacao/rio-doce/rio-doce-saiba-mais, accessed in July 2025.
  » https://www.gov.br/ana/pt-br/sala-de-situacao/rio-doce/rio-doce-saiba-mais
• ³⁸ Gonçalves, J. A. C.; Carneiro, G. A.; Carneiro, G. C. A.; Revista Brasileira de Geografia Física 2020, 13, 1213. [Crossref]
  » Crossref
• ³⁹ United States Environmental Protection Agency (EPA); Method 3051A: Microwave Assisted Acid Digestion of Aqueous Samples and Extracts, part of Test Methods for Evaluating Solid Waste, Physical/Chemical Methods, https://www.epa.gov/esam/us-epa-method-3051a-microwave-assisted-acid-digestion-sediments-sludges-and-oils, accessed in July 2025.
  » https://www.epa.gov/esam/us-epa-method-3051a-microwave-assisted-acid-digestion-sediments-sludges-and-oils
• ⁴⁰ United States Environmental Protection Agency (EPA); Method 6020B: Inductively Coupled Plasma - Mass Spectrometry, part of Test Methods for Evaluating Solid Waste, Physical/Chemical, https://www.epa.gov/esam/epa-method-6020b-sw-846-inductively-coupled-plasma-mass-spectrometry, accessed in July 2025.
  » https://www.epa.gov/esam/epa-method-6020b-sw-846-inductively-coupled-plasma-mass-spectrometry
• ⁴¹ Blanchet, F. G.; Legendre, P.; Borcard, D.; Ecology 2008, 89, 2623. [Crossref]
  » Crossref
• ⁴² Vegan: Community Ecology Package, https://cran.r-project.org/web/packages/vegan/vegan.pdf, accessed in July 2025.
  » https://cran.r-project.org/web/packages/vegan/vegan.pdf
• ⁴³ Dray, S.; Bauman, D.; Blanchet, G.; Borcard, D.; Clappe, S.; Guenard, G.; Jombart, T.; Larocque, G.; Legendre, P.; Madi, N.; Wagner, H. H.; Adespatial: Multivariate Multiscale Spatial Analysis, https://cran.r-project.org/web/packages/adespatial/index.html, accessed in July 2025.
  » https://cran.r-project.org/web/packages/adespatial/index.html
• ⁴⁴ MAPBIOMAS; https://mapbiomasbrsite.s3.amazonaws.com/ATBD_Collection_5_v1.pdf, accessed in July 2025.
  » https://mapbiomasbrsite.s3.amazonaws.com/ATBD_Collection_5_v1.pdf
• ⁴⁵ Souza, Jr. C. M.; Shimbo, J. Z.; Rosa, M. R.; Parente, L. L.; Alencar, A. A.; Rudorff, B. F. T.; Hasenack, H.; Matsumoto, M.; Ferreira, L. G.; Souza-Filho, P. W. M.; de Oliveira, S. W.; Rocha, W. F.; Fonseca, A. V.; Marques, C. B.; Diniz, C. G.; Costa, D.; Monteiro, D.; Rosa, E. R.; Vélez-Martin, E.; Weber, E. J.; Lenti, F. E. B.; Paternost, F. F.; Pareyn, F. G. C.; Siqueira, J. V.; Viera, J. L.; Neto, L. C. F.; Saraiva, M. M.; Sales, M. H.; Salgado, M. P. G.; Vasconcelos, R.; Galano, S.; Mesquita, V. V.; Azevedo, T.; Remote Sens. 2020, 12, 2735. [Crossref]
  » Crossref
• ⁴⁶ Secretaria Estadual de Meio Ambiente e Desenvolvimento Sustentável (SEMAD); https://geoportal.meioambiente.mg.gov.br/, accessed in July 2025.
  » https://geoportal.meioambiente.mg.gov.br/
• ⁴⁷ Empresa Brasileira de Pesquisa Agropecuária Embrapa Solos (Embrapa); Sistema Brasileiro de Classificação de Solos; EMBRAPA: Brasília, 2018. https://www.agroapi.cnptia.embrapa.br/portal/assets/docs/SiBCS-2018-ISBN-9788570358004.pdf, accessed in July 2025.
  » https://www.agroapi.cnptia.embrapa.br/portal/assets/docs/SiBCS-2018-ISBN-9788570358004.pdf
• ⁴⁸ Soil Classification System - USDA; https://passel2.unl.edu/view/lesson/2eafec8dd762/2, accessed in July 2025.
  » https://passel2.unl.edu/view/lesson/2eafec8dd762/2
• ⁴⁹ Karaouzas, I.; Kapetanaki, N.; Mentzafou, A.; Kanellopoulos, T. D.; Skoulikidis, N.; Chemosphere 2021, 263, 128192. [Crossref]
  » Crossref
• ⁵⁰ Instituto Nacional de Metrologia, Qualidade e Tecnologia (INMETRO); Orientação sobre Validação de Métodos Analíticos - DOQ-CGCRE-008, Revisão 09, 2020. https://www.gov.br/cdtn/pt-br/assuntos/documentos-cgcre-abnt-nbr-iso-iec-17025/doq-cgcre-008/view, accessed in July 2025.
  » https://www.gov.br/cdtn/pt-br/assuntos/documentos-cgcre-abnt-nbr-iso-iec-17025/doq-cgcre-008/view
• ⁵¹ Ferreira, C. A.; Palmieri, H. E. L.; Menezes, M. A. B. C.; Auler, L. M. L. A.; Am. J. Water Resour. 2017, 5, 29. [Crossref]
  » Crossref
• ⁵² Brito, M. C. W.; Barbosa, F. A. R.; May, P.; Maroun, C.; Renshaw, J.; Sánchez, L. E.; Kakabadse, Y. ; Abordagens Fonte-Mar e de Paisagem: Integração da Qualidade da Água e Conservação da Biodiversidade na Restauração da Bacia do Rio Doce; IUCN: Gland, 2021. [Crossref]
  » Crossref
• ⁵³ Agência Nacional de Águas e Saneamento Básico (ANA); PIRH Bacia do Rio Doce - Plano Integrado de Recursos Hídricos da Bacia Hidrográfica do Rio Doce, http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/969, accessed in July 2025.
  » http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/969
• ⁵⁴ Agência Nacional de Águas e Saneamento Básico (ANA); ATLAS Esgotos: Despoluição de Bacias Hidrográficas, https://app.powerbi.com/view?r=eyJrIjoiZjA1ZjQwZWUtYmRkYS00YjM0LWFhMjItMTMyOTQ0ND1jNGQyIiwidCI6ImUwYmI0MDEyLTgxMGItNDY5YS04YjRkLTY2N2ZjZDFiYWY4OCJ9, accessed in July 2025.
  » https://app.powerbi.com/view?r=eyJrIjoiZjA1ZjQwZWUtYmRkYS00YjM0LWFhMjItMTMyOTQ0ND1jNGQyIiwidCI6ImUwYmI0MDEyLTgxMGItNDY5YS04YjRkLTY2N2ZjZDFiYWY4OCJ9
• ⁵⁵ Xie, Q.; Li, Z.; Yang, L.; Lv, J.; Jobe, T. O.; Wang, Q.; PLoS One 2015, 10, e0136606. [Crossref]
  » Crossref
• ⁵⁶ Leite, F. P.; Novais, R. F.; Silva, I. R.; Barros, N. F.; Neves, J. C. L.; Medeiros, A. G. B.; Ventrella, M. C.; Villani, E. M. A.; Rev. Bras. Cienc. Solo 2014, 38, 193. [Crossref]
  » Crossref
• ⁵⁷ Instituto Mineiro de Gestão das Água (IGAM); Encarte Especial sobre a Qualidade das Águas do Rio Doce após 2 Anos do Rompimento de Barragem de Fundão - 2015/2017, https://igam.mg.gov.br/w/monitoramento-da-qualidade-das-aguas-superficiais-do-rio-doce-no-estado-de-minas-gerais, accessed in July 2025.
  » https://igam.mg.gov.br/w/monitoramento-da-qualidade-das-aguas-superficiais-do-rio-doce-no-estado-de-minas-gerais
• ⁵⁸ Instituto Mineiro de Gestão das Água (IGAM); Encarte Especial sobre a Qualidade das Águas do Rio Doce após 1 ano do Rompimento de Barragem de Fundão -2015/2016, http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/4379, accessed in July 2025.
  » http://repositorioigam.meioambiente.mg.gov.br/handle/123456789/4379
• ⁵⁹ Instituto Mineiro de Gestão das Água (IGAM); Encarte Especial sobre a Qualidade das Águas do Rio Doce após 4 anos do Rompimento de Barragem de Fundão - 2015/2019, http://www.repositorioigam.meioambiente.mg.gov.br/jspui/handle/123456789/4208, accessed in July 2025.
  » http://www.repositorioigam.meioambiente.mg.gov.br/jspui/handle/123456789/4208
• ⁶⁰ Queiroz, H. M.; Ying, S. C.; Abernathy, M.; Barcellos, D.; Gabriel, F. A.; Otero, X. L.; Nóbrega, G. N.; Bernardino, A. F.; Ferreira, T. O.; Environ. Int. 2021, 146, 106284. [Crossref]
  » Crossref

Edited by

• Editor handled this article:
  Josué Carinhanha Caldas Santos (Associate)

Publication Dates

• Publication in this collection
  15 Sept 2025
• Date of issue
  2025

History

• Received
  31 May 2025
• Accepted
  13 Aug 2025