Application of Multivariate Statistic of U, Th and Pb Concentrations and Pb Isotopic Signatures in the Assessment of Geogenic and Anthropogenic Sources in a U-Mineralized Area

This work presents a Principal Component Analysis (PCA) study using Pbisotope signatures and U, Th and Pb concentrations from groundwater, sediments and rocks (granites and orthogneisses) of the Complex of Lagoa Real (Bahia, Brazil). This area is naturally enriched in U and Th, with the occurrence of Pb derived from the radioactive decay of the elements (U, U and Th) in the form of their stable isotopes Pb, Pb and Pb in addition to the natural isotope Pb. Sampling was carried out in the rainy season (December to January) and the points were selected according to regional hydrology and geology. Thirty samples were analyzed: 12 of groundwater (AP) and 18 of sediments (S). The results show that the use of isotopic ratios allows discrimination between geogenic and anthropogenic samples. This information is not obtained using only the analysis of concentration data. Statistically, the isotopic data of Pb stand out as an efficient tool in the characterization of sources in the scenario investigated, allowing an effective environmental monitoring and a better management of the mining activities.


Introduction
The Uraniferous Province of Lagoa Real (Bahia, Brazil) is an area naturally this element, largely responsible for the economic development of the region. In these areas, the groundwater represents the main source of water for the population, since water scarcity is significant.
According to Licht (2001), most of the natural or anthropic processes occurring on the Earth's surface are characterized by associations of chemical elements (characteristic of the generating process) and geographically associated to their area of occurrence. However, these geochemical signatures may be masked by the superposition of other ones from different sources. Thus, the identifiable geochemical signal, e.g. in waters and sediments, would represent a sum of contributions originated from the natural environmental processes and human action. According to this author, when a complex geochemical signal is interpreted using a multidisciplinary approach, it is possible to decompose it, and, hence, natural or anthropogenic processes can be identified and isolated.
In order to investigate the correlation between a given element and the source that originated it (natural or anthropogenic), works from the 1960s onwards showed the applicability of the methodology of Pb isotope ratios in environmental studies (Nriagu, 1989). Considering that physical or chemical processes do not affect the isotopic composition of Pb in the environment (Doe, 1970;Bollhöfer et al., 1999), the Pb signature tends to be identical to the source that originated it. Therefore, the isotopic signatures of Pb of anthropogenic provenance are distinct from the isotopic signature of natural origin, and can be used as a tracer of the possible sources of Pb in the environment (Bollhöfer et al., 1999;Bollhöfer and Rosman, 2000;Bollhöfer and Rosman, 2001;Chen et al., 2005;Gioia et al., 2006;Gulson et al., 2007;Komárek et al., 2008;Gioia et al., 2010;Vecchia, 2015;Vecchia et al., 2015;Vecchia et al., 2017a and2017b).
In the area of the Gneissic-Granitic Complex, Cordani et al. (1992) obtained data on isotopic composition of Pb in granite rocks of the São Timóteo type (non-deformed granites) and in orthogneisses (deformed granites and associated with uraniferous mineralization). Subsequently, Iyer et al. (1999) demonstrated the applicability of Pb isotopes in the quantitative determination of U and Th mobility in different geological formations present in the São Francisco Craton. These authors studied several geological environments belonging to the São Francisco Craton, among which is the Lagoa Real Gneissic-Granitic Complex, using the isotopic data of Pb published by Cordani et al. (1992).
According to Vecchia (2015), Vecchia et al. (2015) and Vecchia et al. (2017a and2017b), the use of the Pb isotopic tool to study environmentally impacted areas requires knowledge of the Pb isotopic signatures of geographic samples as a determinant factor in the investigative process of environmental matrices.
Therefore, this study uses the Pb isotopic data of geographic samples of the study area obtained by Cordani et al. (1992) along with the U, Th and Pb concentra-  Iyer et al., 1999. The isotopic data of Pb produced by Cordani et al. (1992), containing information on the preserved geological environment, were compared to the isotopic data of Pb from Vecchia et al. (2017b), in order to diagnose influences of geogenic and/or anthropogenic sources in this area.
Statistical analyses allow the specification of the variables that best characterize the geogenic influences coming from the geological context and/or the anthropic activities in the environment. The multivariate statistical methods allow the extraction of useful information capable of reducing the dimensional representation of the data, organizing them in a structure that facilitates the visualization of the whole data set. Principal Component Analysis (PCA) is among the most widely used with this purpose. It compacts the data into a smaller number of variables (principal components) whose variance (and importance) is given by their eigenvalues. These components are mathematical functions that generate scores that have embedded the contributions (and importance) of each variable with its weight (loadings) (Manly, 2005;Mingoti, 2007;Zhao et al., 2011, Cavalcante et al., 2013Sun et al., 2013;Siepak and Sojka, 2017). It is important to emphasize that the results obtained with these techniques should, whenever possible, be confronted with information related to the research context, seeking an association on the possible processes in the study region (Vecchia, 2015;Vecchia et al., 2015;Vecchia et al., 2017a and2017b).
This work presents a statistical study based on the PCA, using Pb isotopic ratios, and U, Th and Pb concentrations from groundwater and sediments in the Lagoa Real Uranium Complex (Brazil). The main objective is to refine the information from the investigations carried out by Vecchia et al. (2017b) by analyzing the contributions of each variable in the diagnosis of possibly impacted areas in the Uraniferous Province area. This work aims to diagnose not only the water quality and sediments in the study area, but also to identify the areas possibly affected by the mining activity.

Study Area
The study area, Southeastern Brazil, is located in the semi-arid region characterized by a marked water deficit. It is considered the main active site of uranium in Brazil and South America. Pb isotopic signature of samples was determined in an area of 1200 km 2 at Lagoa Real Uranium Complex, Bahia State, Brazil ( Figure   1). This complex is geologically identified as a Gneissic-Granitic Complex with ages between of 1.72 and 1.77 Ga (Cordani et al. 1992). Granite type São Timóteo (undeformed granites), orthogneiss (deformed granites and associated U-mineralization) and U-albitites (main hosts of the uranium mineralization) are the principal lithologies of the region and represents the geogenic sources (Vecchia et al., 2017b).
The average annual rainfall is around 700 mm, with high variability in the spatial and temporal distribution (annual seasonality) (Vecchia, 2015). According to Lamego Simões et al. (2003), the climatological and hydrological characteristics, associated to the conformation of the regional relief with flows to Atlantic slope, give rise to a hydrographic network in which intermittent drainage is common ( Figure 1). Complementary hydrological data and geology of the study area can be found in Vecchia (2015).

Sampling
Sampling was carried out according to Vecchia et al. (2017b). Groundwater and sediment sampling points were selected according to regional hydrology and geology, including points at the mining area and points located upstream and downstream thef U anomalies (contents higher than 1500 ppm). The deficit of surface water led to the choice of groundwater and sediment as study matrices. Sampling was carried out in the rainy season (December to January) because of the discontinuous flow of the rivers and the low precipitation level in the region (around 700 mm). The location of sampling was devised aiming to achieve a reliable geographical representation and to ensure a good spatial distribution of the data. Thirty samples were analyzed; 12 of groundwater (AP) and 18 of sediments (S) as shown in Figure 1. The sampling points were georeferenced, using GPS (Global Positioning System).
The sampling and preservation of water samples were carried out according to the Environmental Agency of the State of São Paulo directions-CETESB (2011) and Veridiana et al. (2008). More details in Vecchia et al. (2017b).
Approximately 1 kg of sediments samples, representative of the regional lithologies (Figure 1), were collected in depth from 50 cm to 80 cm, using a cylindrical collector. This material was transferred to low-density polyethylene (LDPE; Nalgon  ) containers. Subsequently, these samples were dried, disaggregated, homogenized and sieved. Fractions < 250 Mesh Tyler were sent to isotopic Pb and chemical analysis.

Pb Isotopic Geochemical Analytical Procedure
Pb isotopic ratios of groundwater and sediments were determined through

Multivariate Analysis
Data were submitted to a Principal Component Analysis aiming to establish relations between U, Th and Pb contents and Pb isotopic ratios (

Multivariate Analysis
Data (Tables 1-3    2) PCA using Pb isotopic ratios in sediments, groundwater and orthogneiss samples The same procedure was applied to the same samples, removing U, Th and Pb concentrations, i.e., working only with the Pb isotopic ratios as variables. Table   2 shows the loading values for both components. Concerning the isotopic ratios there is no difference between the results obtained with the two approaches, i.e., calculations with and without U, Th and Pb concentrations. Figure 3 and Figure   4 reflect this result. The distributions of samples in both diagrams are the same.    3) PCA using Pb isotopic ratios data for sediments, groundwater and São Timóteo Granite samples.
For this matrix, only data of Pb isotopic ratios were used. They refer to the samples of São Timóteo type granite, groundwater and sediments.  According to Vecchia et al. (2017b) there is a great variation of Pb isotopic

Conclusion
One of the major vulnerabilities of the nuclear area and/or mining activities to public opinion refers to frequent suspicions that nuclear and/or mining industry always generate(s) large environmental impacts, especially contamination of groundwater. These kinds of worries can find legal resonance even without the necessary pre-operational background data, and most of the decisions consider the analytical results of radionuclide concentrations from the operational phase as the main incriminating technical argument. The present study showed that the determination of the concentration of the contaminants is not enough to characterize environmental contamination. It also shows that the most effective analytical method to elucidate the role of anthropogenic interferences in the environment is the determination of isotopic ratios. It can even be used in environments with high concentrations of radionuclides.
This tool becomes even more relevant in regions where pre-operational data are scarce and incomplete, if not inexistent, or even in regions, such as Caetité (Bahia), where water sampling is very difficult due to low water availability. In addition, it has been proved that the Pb isotopic tool, besides being used in any type of environmental matrices, is still promising in the investigation of industrial environmental impacts from mining as well as from physical, chemical and thermal processes where the generation of NORM (Natural Occurring Radioactive Material) wastes is related.