Journal of South American Earth Sciences xxx (xxxx) xxx

Contents lists available at ScienceDirect

Journal of South American Earth Sciences

journal homepage: www.elsevier.com/locate/jsames

Natural Hydrogeochemical Background Levels in the Carste Lagoa Santa

Protection Unit, Minas Gerais, Brazil
Frederico Aragão *, Leila Nunes Menegasse Velásquez, Paulo Galvão, Luísa Costa Vieira
Federal University of Minas Gerais, Institute of Geosciences, Pampulha Campus, Belo Horizonte, MG, 31275-035, Brazil

A R T I C L E I N F O A B S T R A C T

Keywords: The identification of hydrogeochemical anomalies in karst water resources represents a crucial step in Brazil’s
Natural background level water management plans. Accelerated urban and industrial development occurs in the region that includes the
Carste lagoa santa protection unit Carste Lagoa Santa Protection Unit, State of Minas Gerais, which may negatively impact its water quality. This
Hydrogeochemistry
paper establishes Natural Background Levels (NBL) of waters collected from deep tubular wells and springs for
Karst aquifer
Geogenic anomalies
Si4+, Ca2+, Mg2+, Na+, K+, HCO−3 , SO24, Cl− to understand sources of possible geogenic and anthropogenic
anomalies. Areas with surpassed thresholds were compared with geological, hydrogeochemical, and land use
data. Every parameter, except for chloride, showed elevated concentrations due to lithologic origin. The NBLs for
waters of wells and springs are, respectively, 17.9 and 12.7 mg/L (silica), 97.0 and 91.5 mg/L (calcium), 4.9 and
2.9 mg/L (magnesium), 6.2 and 3.8 mg/L (sodium), 1.0 and 0.3 mg/L (potassium), 285.5 and 254.6 mg/L
(bicarbonate), 9.4 and 13.6 mg/L (sulfate). Chloride, analyzed only in groundwater, presented 6.7 mg/L indi­
cating a possible influence of upstream urban areas of Lagoa Santa and Pedro Leopoldo municipalities. Volatile
and Semivolatile Organic Compounds were also analyzed, but all of them were below the detection limit. This
work is important to comprehend the NBL of major elements, reveal geogenic anomalies, and distinguish car­
bonate aquifer units from the impure ones.

Author contribution
Frederico Aragão, Conceptualization, Methodology, Software,
Formal analysis, Writing - original draft, Writing - review & editing,
Leila Velásquez, Supervision, Funding acquisition, Formal analysis,
Paulo Galvão, Supervision, Writing - review & editing, Formal analysis,
Luisa Vieira, Writing - review & editing.
1. Introduction
Groundwater quality plays a crucial role in environmental manage­
ment studies. Populations often rely on this resource and, therefore,
proper monitoring and protection are required. Karst aquifers are
important groundwater sources, where it is estimated that more than
20–25% of the world depends on water from these aquifers (Ford and
Williams, 2007; Chen et al., 2017), and by 2025 it is expected that this
percentage will rise considerably to approximately 80% (Forti, 2002).
Karst aquifers have unique hydrogeological characteristics, as they
are complex, more anisotropic, and heterogeneous, compared to porous
media and fractured aquifers. These features result, for instance, in
double types of recharges (concentrated or diffuse – autogenic and
allogenic), and faster groundwater fluxes through conduits. Flow ve­
locities can reach high values, such as up to 2.66 m/d in West Virginia
(Jones, 1977) or even 5.18 m/d in Herzegovina (Goldscheider and
Drew, 2007), with a heterogeneous network allowing the integration of
flow path from the surface to spring (White 1988; European Commis­
sion, 1995; Bakalowicz, 2005; Ford and Williams, 2007; Mudarra and
Andreo, 2011).
The heterogeneity constitutes a complex flow behavior, conse­
quently producing a highly variable hydrogeochemistry. These varia­
tions depend on rock mineralogy, residence-time, and reactive
processes, as well as flux conditions and permeability properties. A
highly mineralized groundwater composition, result of a paleo-recharge
and circulation processes, has been studied, whereas other studies
pointed weakly mineralized composition coupled with fast transfer
conditions (Cook et al., 2005; Bucher et al., 2008; Bucher and Stober,
2010; Armandine et al., 2014).
The protection of groundwater in karst aquifers can be achieved by

* Corresponding author. Federal University of Minas Gerais, Belo Horizonte, Minas Gerais, Brazil.
E-mail address: aragao.ambiental@gmail.com (F. Aragão).

https://doi.org/10.1016/j.jsames.2020.102985
Received 7 July 2020; Received in revised form 9 October 2020; Accepted 22 October 2020
Available online 5 November 2020
0895-9811/© 2020 Elsevier Ltd. All rights reserved.

Please cite this article as: Frederico Aragão, Journal of South American Earth Sciences, https://doi.org/10.1016/j.jsames.2020.102985
F. Aragão et al.
Fig. 1. CLSPU Land use map extracted from Correa (2018) complemented with industrial and urban activities as well as the location of the study area in
South America.
intrinsic mapping vulnerability to contamination and the study of
contaminant load generation (Doerfliger et al., 1999; Pereira et al.,
2019). The delimitation of protection and catchment areas of ground­
water supply sources, the evaluation of the contamination risk, and the
investigation of the activities that generate contaminants should be
addressed. Finally, groundwater protection measures, such as control of
effluents treatment, land use, and urban planning should be performed
to effectively protect the resources (Foster et al., 2006).
A method that can be used to support the protection of groundwater
resources in karst terrains is the determination of the Natural Back­
ground Level (NBL) for hydrogeochemical parameters. This concept
corresponds to the range of a substance or element concentration ob­
tained from the interaction between groundwater, surface water, litho­
logic, biologic, and atmospheric factors, under the little or absent
anthropogenic impact. The possible linkage between aquifers, rain, and
small anthropogenic outputs might influence those values; thus, spatial
variations are expected (ISPRA, 2009). NBL helps investigate whether a
value for an element or substance corresponds to a geogenic origin or
not (Hawkes and Webb, 1962; Reinmann et al., 2005; Edmunds and
Shand 2008; Hinsby and Rasmussen, 2008). In 2006, a method for
determining the NBL of a substance was developed by the European
community, in a project named "Background Criteria for the Identifi­
cation of Groundwater Thresholds" (BRIDGE 2006, 2009).
In Brazil, 52% of the 5,570 Brazilian municipalities are totally (36%)
or partially (16%) supplied by groundwater (ANA, 2012). However,
even though karst areas in the country comprise between 5% and 7% of
the national territory (425,000–600,000 km2 – Auler et al., 2001), few
studies regarding natural background levels exist in karst aquifers.
Lucon et al. (2018) presented NBL values for calcium, magnesium, so­
dium, potassium, bicarbonate, sulfate, chlorine, nitrate, and phosphate
for the São Miguel watershed, state of Minas Gerais, pointing to
anthropogenic impacts for NO−3 , PO3− 4 , and Cl .
−
The study area includes the Carste Lagoa Santa Protection Unit
(CLSPU), with karst geomorphology features composed of springs,
2
Journal of South American Earth Sciences xxx (xxxx) xxx
swallow holes, caves, sinkholes, shafts, and poljes. Geologically, the area
is in low-grade metamorphism, and it is constituted of the Neo­
proterozoic Sete Lagoas Formation (impure limestones of the Pedro
Leopoldo Member and pure limestones of the Lagoa Santa Member), and
Serra de Santa Helena Formation (metapelites). An uprising urban and
industrial development is taking place in this zone, especially in
southwest and west borders, as the area is 52 km distant to the capital
Belo Horizonte, the state of Minas Gerais capital. Industries, agricultural
and urban settlements indicate risks to surface and groundwater quality
since an appropriate treatment is not always carried out in effluents and
waste (Tayer, 2016). As for volatile and semivolatile organic compounds
(VOCs and SVOCs), since they represent substances that, once into
groundwater, are persistent and can move to supply-wells, constitute
public health substances that are worth monitoring (National Research
Council, 2001).
Due to these features, this paper proposes NBL values for inorganic
elements in the groundwaters of the CLSPU to indicate if the surpassed
values indicate geogenic anomalies or possible anthropogenic inter­
vention. A hydrogeochemical analysis is also carried, supported by the
NBLs, to help distinguish the aquifer units. The NBLs were then calcu­
lated for samples collected from deep tubular wells and springs for
calcium, bicarbonate, magnesium, silica, sodium, potassium, and sulfate
and only from deep tubular wells for chloride.
2. Site description
The study area (Fig. 1) is in the state of Minas Gerais, the southeast
region of Brazil, 52 km distant from Belo Horizonte, the state capital. It
includes the CLSPU completely and partially seven surrounding mu­
nicipalities, in a total of 505 km2. The zone is inserted in the upper São
Francisco river basin and the Velhas river subbasin, with an altitude
between 620 and 910 m above sea level (Tayer and Velásquez, 2017).
The most common activities in the region that cause environmental
impacts are fuel stations, steel industry, cemeteries, limestone mining,
F. Aragão et al.
Fig. 2. Geological map with the allocation of deep tubular wells in the respective lithologies in which the waters were collected.
intensive agriculture, and urban sites with poor sanitation structures
(Tayer, 2016). The land use map (Correa, 2018) indicates considerable
areas where, notably, agricultural activities and urban settlements are
frequent on Fig. 1 (the coordinate system used for all figures was Uni­
versal Transverse Mercator (UTM) projection, zone 23S, datum WGS
1984, with units in meters).
Geologically, the area is in the São Francisco Craton, in which
Neoproterozoic pelite-carbonate rocks from Sete Lagoas and Serra de
Santa Helena formations compose the Bambuí Group. The Belo Hori­
zonte complex, which forms the Archaean basement, comprises gneiss,
migmatites, and granitoids. Moreover, the Sete Lagoas Formation is
segmented into two members: Pedro Leopoldo, at the bottom, repre­
sented by fine limestones, dolomites, marlstones, and pelites, showing a
calcium carbonate composition between 60% and 90%; and Lagoa Santa
Member, at the upper portion, constituted by purer medium-grained
black limestones (calcium carbonate content higher than 90%) and
with a substantial karstification level. The Serra de Santa Helena For­
mation covers the Sete Lagoas Formation, and it is composed of argillite,
siltstone, slate, and marble. The Bambuí Group is overlaid by Cenozoic
unconsolidated sediments, in which drainages, watercourses meanders,
3
Journal of South American Earth Sciences xxx (xxxx) xxx
and low topographic levels are located, being segmented in detritus
coverage, alluvial terraces, and alluvium (Ribeiro et al., 2003; Tuller
et al., 2010; Galvão et al., 2016), which is depicted in Fig. 2.
Hydrogeologically, karst zones were formed over limestone rocks of
the Sete Lagoas Formation, having unconfined conditions when these
rocks outcrop, and being semi-confined when overlying metapelites
from the Serra de Santa Helena Formation act as an aquiclude (Mourão
et al., 2001). The Lagoa Santa Member constitutes the more karstified
zones, while the Pedro Leopoldo Member is constituted of conduits
developed in sub-horizontal bedding planes. As these members are hy­
draulically connected through subvertical fractures and complex
conduit networks, it is constituted the Sete Lagoas Karst System (SLKA –
Pessoa 2005; Galvão 2015; Tayer and Velásquez, 2017). In this system,
karst conduits generally contain water of satisfactory quality for human
consumption, despite the high hardness (Pessoa 2005; Tayer and
Velásquez, 2017). Underneath the SLKA, the basement Belo Horizonte
Complex presents groundwater in faults and fractures, acting as a frac­
tured aquifer.
The CLSPU and its surroundings present two types of recharge: 1)
autogenic, where karst recharges itself from rainwater through
F. Aragão et al.
Fig. 3. Karst features and drainage system of the study area.
sinkholes, poljes, uvalas, and cave entrances, where the Sete Lagoas
Formation outcrops; and 2) allogenic, where the overlying Serra de
Santa Helena’s rocks and Cenozoic unconsolidated sediments occur,
taking place when adjacent non-karst areas recharge the karst aquifer.
There is also diffuse recharge through thick lateritic soils resulting in
attenuation of pollutants (Aragao et al., 2020).
The waters are hydrochemically classified as predominantly
calcium-bicarbonate and, subordinately, as mixed bicarbonate, referring
to the granite and sodic, associated with the colluvium. There is a hor­
izontal compartmentation tendency, demonstrated by increased Si4+
concentration from NW to SE, also extending to NE and followed by
Mg2+ concentrations. This pattern is developed from the Lagoa Santa
Member pure limestones to the impure rocks of the Pedro Leopoldo
Member, showing an increase of dolomite, mica, and chlorite in the east
of the region. The Pedro Leopoldo Member’s waters can be divided into
4 hydrochemical types, mainly according to Ca2+ and Si4+ concentra­
tions and, secondly, to Mg2+ concentrations (types i, ii, iii, iv).
4
Journal of South American Earth Sciences xxx (xxxx) xxx
Moreover, they can be associated with different classifications of rocks
through lithochemistry, as proposed by Vieira (2018).
The endokarst is mainly represented by grottos, whereas the exokarst
features are illustrated as poljes, shafts, sinkholes, caves, ponors,
springs, and uvalas (Ribeiro et al., 2003). The discharge areas show a
substantial level of dissection below 700 m, presenting residual karstic
features, substantial uvala systems, and relatively flat terrain. The
sinkholes, caves, dolines, and the main drainage system, having the
Velhas river as the most important one, are illustrated in Fig. 3.
3. Material and methods
To understand the carbonate and the pelitic lithologies influences
over the water quality and the possibility of geogenic anomalies, sam­
ples from tubular wells and springs were collected, as illustrated by
Fig. 2. A stainless-steel bucket was used in all samplings, which was
rinsed three times using the collected water.
F. Aragão et al. Journal of South American Earth Sciences xxx (xxxx) xxx

Table 1 laboratory on the same day, where they were kept at a temperature of
Distribution and descriptive statistics for the collected deep tubular wells. 4 ◦ C for a maximum of 24 h before the analyses, according to the
Parameter Distribution AD P- n Mean Std. Median Sampling and Water Preservation Guide of São Paulo’s State Sanitation
value final Dev Technology Company (CETESB, 2011). The sampling was carried out at
Ca2+ Logistic 0.2 >0.3 77 67.6 26 65.7 the possible upstream point to obtain the most preserved physico­
HCO3- Logistic 0.6 0.1 75 214.2 58.8 211.1 chemical characteristics for springs.
Si4+ 3-Parameter 0.6 0.1 77 13.9 6.4 11.7 All the samplings were carried out between September 2017 and
Weibull April 2018, during the rainy season. The samples were analyzed at the
Mg2+ 3-Parameter 0.2 77 2.7 1.8 2.3
SGS Geosol Laboratórios Ltda., in Belo Horizonte. Calcium, magnesium,
>0.5
Weibull
Na+ 3-Parameter 0.1 >0.5 49 3.1 3 2.1 potassium, sodium, and phosphate were determined by metals by
Weibull plasma emission spectroscopy; chloride and sulfate by ions chromatog­
K Loglogistic 0.6 0.1 43 0.7 0.5 0.3 raphy; bicarbonate alkalinity, and silica by the molibdosilicate method.
+

Cl− 3-Parameter 0.3 22 3.8 4.1 2.3

>0.5
Regarding the organic parameters, the VOCs and SVOCs were deter­
Weibull
SO2- 3-Parameter 0.5 0.3 38 4.8 6.1 2.8 mined by gas chromatography/mass Spectrometry, by the Validated
4
Weibull Test Method 8260C (EPA, 2006) and 8270D (EPA, 2014), respectively.
Finally, the values obtained for Si4+, Ca2+, Mg2+, Na+, K+, HCO−3 ,
2−
SO4 , Cl− were compared with the Brazilian Federal drinking standards,
Table 2 found in BRASIL (2017). The laboratory analyses included 96 organic
Distributions and descriptive statistics for the collected springs. contaminants, including benzene, ethylbenzene, toluene, xylene, and
pesticides, carried out for 16 springs and 22 deep tubular wells.
Parameter Distribution AD P- n Mean Std. Median
value final Dev

Ca2+ Smallest Ext. 0.6 0.1 20 71.4 21.7 75.8 3.1. Non-detect treatment
Value
HCO−3 Smallest Ext. 0.7 0.1 20 206.1 53.5 219 In samples with values below the detection limit (DL), it is recom­
Value mended by ISPRA (2009) to convert them to DL itself to identify a
Si4+ 3-Parameter 0.5 0.2 21 11.1 8.6 8.4
Weibull
representative value, otherwise, it will be characterized as "< DL", and
Mg 2+
Largest Ext. 0.3 >0.3 21 1.7 0.9 1.5 therefore, it would be impossible to manipulate them mathematically.
Value For instance, two magnesium samples in the Sete Lagoas Formation
Na Exponential 1.2 0.7 21 1.6 2.2 0.9 analysis were considered with an exact value of 0.25 mg/L. Addition­
+

K+ 3-Parameter 0.6 0.1 21 0.4 0.2 0.3

Weibull
SO2- Exponential 1.1 0.1 21 5.9 7.1
4
Only deep tubular wells with an average depth of 90 m extracting
water from Sete Lagoas and Serra de Santa Helena formations were
considered for groundwater. 26 wells from the Lagoa Santa Member, 43
from the Pedro Leopoldo Member (total of 69 samples in the Sete Lagoa
Karst Aquifer), 3 from the Serra de Santa Helena Formation, and 5 from
a mix between Lagoa Santa Member and Serra de Santa Helena For­
mation were sampled. In all wells, samples were collected after the
previous pumping for a minimum period of 10 min to purge and restore
original hydrochemical conditions and to avoid stagnant water. As the
wells generally have diameters of 4 and 6′′ , average depths of 90 m, and
pumps extracting flow rates of 20 m3/h (De Paula 2019), the adopted
time to purge was considered sufficient.
As for the springs, 21 points were sampled: 13 of them from the
Lagoa Santa Member, 5 from the Lagoa Santa Member, 2 from the Belo
Horizonte complex basement, and 1 from lateritic Cenozoic unconsoli­
dated sediments. Previously to each sampling, the plastic bottles used to
store sampled water were rinsed three times with the sample to equili­
brate sampling equipment with the sample environment and to elimi­
nate residues that might be present (USGS, 2006). Straight after
sampling, the bottles were placed in a cooler and then taken to the
ally, in order to avoid the uncertainty, which is inherent from the usage
2.8 of those type of data, it is suggested to use this constant value exclusively
when the number of DL samples constitute a maximum of 15% of the
entire dataset (EPA, 2000).
3.2. NBL determination, descriptive statistics, and maps elaboration
The method used to assess the Natural Background Level was pri­
marily suggested by APAT-ISS (2006), and ISPRA (2009), in which the
NBL corresponds to the 90th percentile of the probability distribution
plot. This technique is also found in Tedd et al. (2017), Lucon et al.
(2018, 2020) and Aragão et al. (2020), both of which applied in karst
environments and thus valuing the present method. For that purpose, it
was initially necessary to determine the best statistical distribution for
each parameter, which was performed through the software Minitab®,
version 17. The ’Individual Distribution Identification’ tool compares 14
statistical distributions by the Anderson Darling (AD) test, that provides
the p and the AD values. The combination of a more significant p-value
(with a minimum acceptable value of 5%) coupled with a smaller AD
indicates the best fit, as suggested by Minitab (2018).
The mean, median, and standard deviation were also calculated, and
the maps showing the areas in which the NBLs were surpassed were
produced. To obtain these areas, the Radial Basis interpolation method
(Completely Regularized Spline function) was used, since it is accurate,

Table 3
NBLs, maximum values, drinking standards and detection limits.
Parameter NBL 90 Deep tubular Max. Value Prob. Distr. Deep NBL 90 Springs Max. Value Prob. Distr. Brazilian drinking water standard Detection Limit
wells (mg/L) tubular wells (mg/L) (mg/L) Springs (mg/L) (mg/L) (BRASIL 2017) (mg/L)

Ca2+ 97 162.2 91.5 101 – 0.3

HCO−3 282.5 428.5 254.6 277.5 – 6
Si4+ 17.9 39.5 12.7 29.2 – –
Mg2+ 4.9 12.1 2.9 4.6 – 0.3
Na+ 6.2 22.2 3.8 7.5 200 0.5
K+ 1 2.1 0.3 1 – 0.3
Cl− 6.7 17.9 – – 250 0.3
SO2-4 9.4 22.6 13.6 27.3 250 1.2

5
F. Aragão et al. Journal of South American Earth Sciences xxx (xxxx) xxx

Fig. 4. Areas with surpassed NBLs and probability plots for calcium collected from deep tubular wells and springs (A) and for bicarbonate (B).

flexible and produces "the best overall interpretations of most data sets"
when compared to other interpolation techniques (Golden Software
2017; Gunarathna et al., 2016). The value of 2 km buffer was chosen for
the boundaries in all maps to provide a security margin to more assertive
interpretations of the analyses (Persson et al., 2010; Ajonye et al., 2016).
3.3. Groundwater analysis
As for the cationic parameters, calcium was the only parameter that
did not present any samples with concentration below the detection
limit. The tool Individual Distribution Identification suggested Logistic

6
F. Aragão et al.
Fig. 5. Areas with surpassed NBLs and probability plots for silica collected from deep tubular wells and springs (A) and for magnesium (B).
distribution as the one with the best fit since its p-value (>25%) and AD
value (0.3) was the best combination among the 14 distributions. Silica
and calcium showed an AD of 0.6 and a p-value of 13.5%, under 3-
Parameter Weibull distribution. Bicarbonate presented a p-value of
8.1%, and an AD of 0.6 was attained, all of them within the Small
Extreme Value distribution.
Moreover, magnesium showed a satisfactory fit, with p > 0.5 and AD
= 0.2, under the 3-Parameter Weibull distribution, whereas sodium held
an AD of 0.1 and a p-value of >50%, within 3-Parameter Weibull dis­
tribution. Potassium, under the Loglogistic distribution, showed AD and
p-values of 0.6 and 7.8%, respectively.
As for the anions, chloride presented AD and p-values of 0.3 and >
50%, under the 3-Parameter Weibull distribution. For sulfate, it held an
AD and p of 0.5 and 27.5%, also under the 3-Parameter distribution.
Table 1 summarizes those results and the descriptive statistics, including
the mean, standard deviation, and median.
7
Journal of South American Earth Sciences xxx (xxxx) xxx
3.4. Spring analysis
As for the spring cationic parameters, calcium and bicarbonate ob­
tained the Smallest Extreme Value distribution. Ca+2 presented an AD of
0.6 and a p-value of 9.5%, while HCO−3 showed 0.7 and 7%, respectively.
Silica and magnesium presented 0.5 and 19.4% and 0.3 and > 25% for
AD and p-value, under 3-Parameter Weibull and Largest Extreme Value,
respectively. Additionally, the AD was 1.2 and the p-value of 6.9% for
sodium under Exponential distribution, whereas potassium showed a 3-
Parameter Weibull distribution. Finally, as for the anionic parameters,
sulfate presented an AD of 1.1 and a p-value of 8.7%, under the Expo­
nential distribution. However, it was not feasible to perform the chloride
assessment, as there were not enough samples to attain statistical sig­
nificance. Table 2 summarizes the analysis as well as the descriptive
statistics.
F. Aragão et al.
Fig. 6. Areas with surpassed NBLs and probability plots for sodium collected from deep tubular wells and springs (A) and for potassium (B).
4. Results
4.1. NBLs – deep tubular wells and springs
Table 3 summarizes the outcomes for deep tubular wells and springs,
coupled with the maximum values found and the potability drinking
standards (BRASIL, 2017) when defined. The maps that represent areas
where the Natural Background Levels are surpassed for each parameter,
both for deep tubular wells and springs, and the probability plots indi­
cating the statistical distribution and 90th percentile, can be seen in
Figs. 4–7. The organic contaminants’ concentrations, including benzene,
toluene, ethylbenzene, xylene, pesticides, and herbicides, were all below
the quantification limit; therefore, no further considerations were made
them.
4.1.1. Calcium and bicarbonate
Fig. 4A illustrates two main areas in the central-north portion where
the NBL of 97.0 mg/L was surpassed for calcium in waters collected in
8
Journal of South American Earth Sciences xxx (xxxx) xxx
deep tubular wells, while waters from springs the NBLs surpassed
exclusively in the central region, due to the influence of the Lagoa Santa
Member. Furthermore, as for the HCO−3 deep tubular wells, there were
not any expressive areas showing concentration higher than 282.5 mg/
L, whereas for springs, however, a key area with values higher than
254.6 mg/L is observed in the central area, coincident with the patterns
found for calcium.
4.1.2. Silica and magnesium
Fig. 5A illustrates, for silica, areas in the east and smaller portions in
the north and the south for the waters collected from wells. As for the
springs, there are two expressive areas with concentrations above 12.7
mg/L, covering the north, east, south, and southwest, coincident with
the Pedro Leopoldo Member, the Serra de Santa Helena Formation, and
the granitic complex. Mg2+, however, did not show any expressive areas
with concentrations above the calculated natural background levels of
4.9 mg/L for wells and 2.9 mg/L for springs, respectively.
F. Aragão et al.
Fig. 7. Areas with surpassed NBLs and probability plots for sulfate collected from deep tubular wells and springs (A) and for chloride only for water collected from
deep tubular wells (B).
4.1.3. Sodium and potassium
As for sodium from deep tubular wells, Fig. 6A shows no areas with
high NBLs, whereas for springs, there is an area covering the southwest
over the granitic complex because of just one sample. Fig. 6B depicts
only one considerable area in the east for potassium, coupled with three
isolated deep tubular wells in the north, west, and east, which presented
concentrations above 1.0 mg/L. Concerning the springs, potassium is the
parameter with the most extensive areas, where the NBL of 0.3 mg/L is
exceeded, with only the central and east portions below this threshold.
4.1.4. Sulfate and chloride
Fig. 7A shows a coincident area in the center, coupled with two
isolated points where the concentration of 9.4 mg/L is surpassed.
Additionally, Fig. 7B illustrates only one area in the central part, influ­
enced by three springs in the Lagoa Santa Member with NBL higher than
13.6 mg/L. Due to a lack of statistical representativity, it was impossible
to create a map for chloride in spring samples; there were not enough
points, and thus, a satisfactory p-value was not reached. However, the
9
Journal of South American Earth Sciences xxx (xxxx) xxx
map for deep tubular wells was accomplished, and it exhibits a single
area in the west, influenced by three deep tubular wells.
5. Discussion
5.1. Springs
The pattern observed for calcium and bicarbonate is coherent with
the described karst geology. The springs, where the NBL is surpassed, are
deriving water from the Lagoa Santa Member, known for its purer car­
bonate composition, which contributes to the highest Ca2+ and HCO3−
concentrations.
The higher concentrations of silica occur due to the granitic lithology
of the Belo Horizonte Complex and the Pedro Leopoldo Member. In the
east portion, the Pedro Leopoldo Member rocks hold a higher intensity
of pelitic intercalation between carbonate units compared to the other
outcrops, according to fieldworks and lithochemical observations pre­
sented in Vieira et al. (2018). This characteristic influences the chemical
F. Aragão et al.
Fig. 8. Stiff diagrams for the sampled springs (Vieira et al., 2020).
composition of the springs in the east of the area and explains the
elevated Si4+ concentrations.
Likewise, the area where magnesium shows surpassed NBL has its
origin also in the pelitic composition of the Pedro Leopoldo Member,
within its magnesian face. The magnesium concentration is explained by
the dissolution of mica, chlorite, and dolomite, the last one being a
mineral highly present in the limestones of the Pedro Leopoldo Member,
as indicated by lithochemistry data and the X-Ray diffraction analysis
presented in Vieira (2018).
Sodium, on the other hand, is associated with the granitic basement.
However, it is noteworthy that it holds a significant relevance regarding
the distinction of waters from Pedro Leopoldo and Lagoa Santa mem­
bers, as the pelitic intercalations contribute to higher levels of this
parameter in the shallow aquifer circulations.
As for potassium, elevated values were found in almost the entire
region. There is an indication of anthropogenic influence through fer­
tilizer utilization in these values (Correa, 2018) since the surpassed
areas coincide with the agricultural use areas indicated by the land-use
map (Fig. 1). Additionally, these values might occur due to the soil’s
influence, which can release potassium into the groundwater system.
Finally, sulfate shows its pattern possibly because of sulfides oxida­
tion within the rocks, especially pyrite, which is predominantly found in
the limestones of the Lagoa Santa Member and the purer portions of
Pedro Leopoldo Member rocks, as described by Ribeiro et al. (2003),
Tuller et al. (2010), Galvão et al. (2016).
5.2. Deep tubular wells
The patterns observed for calcium and bicarbonate are correlated
and indicate the most karstified areas, with an average karst thickness of
200 m (De Paula 2019). The NBLs are surpassed due to deep tubular
wells exclusively located in the Lagoa Santa Member of pure carbonate
composition, indicating a geogenic anomaly.
As for silica and magnesium, the areas where the NBL is surpassed
correspond to east limits of the area, close to the Velhas river. These
waters belong to the Pedro Leopoldo Member, and the higher concen­
trations of these elements can be explained by pelitic intercalations and
the presence of dolomite. The distribution of Si4+, coupled with Ca2+
10
Journal of South American Earth Sciences xxx (xxxx) xxx
and HCO3− and, in a lesser extent, with Mg2+, enables the identification
of four hydrochemical types in the Pedro Leopoldo Member related to
different lithofacies (Vieira, 2018): type i, with high concentrations of
Ca2+ and HCO−3 , intermediate Mg2+ and low Si4+; type ii, with high
concentrations of Ca2+, HCO−3 and Mg2+ and intermediate of Si4+; type
iii, with low Ca2+ and HCO−3 concentrations, high Si4+, and intermediate
Mg2+; and type iv, with intermediate and low Ca2+ and HCO−3 concen­
trations, low Si4+ and intermediate Mg2+. Silica plays a remarkable role
in the aquifers’ horizontal compartmentation as its concentration in­
creases from NW to SE, also extending to NE and followed by Mg+2
concentrations. Additionally, Si4+ values increase with depth, which is
demonstrated by the hydrochemical types of the Pedro Leopoldo
Member, indicating vertical compartmentation as well (Vieira, 2018).
On the other hand, magnesium exceeds NBL in 4.9 mg/L because of
the impure typology of Pedro Leopoldo Member (group ii, correspond­
ing to dolomitic limestones), with all the three deep tubular wells
extracting water from it.
Potassium indicates pelitic intercalations present in the southeast of
the Pedro Leopoldo Member, with a single well in the north corre­
sponding to the pelite of the Serra de Santa Helena Formation. As for
sodium, there are two deep tubular wells with high potassium concen­
trations, indicating the influence of feldspar and mica dissolution in the
Pedro Leopoldo Member.
Moreover, for chloride, the west’s area and the point in the south had
NBL surpassed by 6.7 mg/L due to water collection of deep tubular wells
from the Pedro Leopoldo Member. However, they represent limited
spots, and this surpassing happened, possibly because of upstream urban
interference in the west of the study area (Fig. 1).
As for sulfate, the central area with the exceeded NBL occurs because
of three deep tubular wells extracting water from the Lagoa Santa
Member, which receives influence from the oxidation of sulfide
disseminated on its rocks (Ribeiro et al., 2003; Tuller et al., 2010; Galvão
et al., 2016).
Lucon et al. (2018) determined Natural Background Levels for cal­
cium, magnesium, sodium, potassium, bicarbonate, sulfate, and chloride
at the São Miguel watershed, in a similar hydrogeological context, 260
km southwest of the CLSPU, but through a different approach. There was
no indication of the probability plots’ statistical distributions, whereas
F. Aragão et al.
Fig. 9. Stiff diagrams for the sampled deep tubular wells (Vieira et al., 2020).
the present study addresses this topic. Additionally, Aragao et al. (2020)
also performed a slightly different NBL analysis for the CLSPU, with no 2
km buffer, therefore with a different number of samples. The studied
parameters were chloride, nitrate, sulfate, and phosphate, and the NBLs
were proposed to validate the assessment of intrinsic vulnerability to the
contamination in the region. As for Cl− and SO2− 4 , different NBLs (2.4
mg/L and 5.3 mg/L) were obtained compared to the present study (6.7
mg/L and 9.4 mg/L), and NO−3 and PO3− 4 had NBLs of 1.0 mg/L and 0.1
mg/L, respectively. Galvão et al. (2017) also showed similar values in
Sete Lagoas, in the southwest of the study area, thus corroborating this
research.
Thus, the surpassing of NBLs is satisfactorily explained by the geo­
genic influence for all the parameters, except chloride. The pure
11
Journal of South American Earth Sciences xxx (xxxx) xxx
carbonate composition of the Lagoa Santa Member is expressed by high
levels of calcium and bicarbonate, while the pelitic intercalations mainly
found in the Pedro Leopoldo Member point to the anomalies found for
silica, magnesium, potassium, and sodium. On the other hand, sulfate
occurs due to the oxidation of sulfide minerals (Ribeiro et al., 2003;
Tuller et al., 2010; Galvão et al., 2016).
Stiff diagrams presented at Vieira et al. (2020) for the sampled
springs and the wells can be seen in Figs. 8 and 9, respectively, illus­
trating the predominance of Ca2+ and HCO−3 in both cases. As for the
springs, the differentiation between waters from Pedro Leopoldo and
Lagoa Santa Members is subtle, due to the fact the waters from the Pedro
Leopoldo Member are associated with purer portions of the rocks.
The deep tubular wells’ Stiff diagrams made it possible to identify
F. Aragão et al. Journal of South American Earth Sciences xxx (xxxx) xxx

Fig. 10. Hydrochemical map created by Vieira (2018), from Si4+ concentrations, showing the 477 Stiff representative composition in the aquifer units. On the Pedro
Leopoldo Member four hydric types were determined from isopleth and concentration maps of Ca2+ and Si4+: (i) High Ca2+ and HCO3- and low Si4+ concen­
trations (ii) High Ca2+, HCO3- and Mg2+ and intermediate Si4+ concentrations (iii) Low Ca2+ and HCO3, elevated Si4+ and intermediate Mg2+ concentrations
(iv) Intermediate and low Ca2+ and HCO3- and low Si4+ concentrations.

seven aquifer units’ hydrochemical facies previously identified. The
mixed waters between the Lagoa Santa Member and Serra de Santa
Helena Formation, the waters only from Serra de Santa Helena, and a
large part of Pedro Leopoldo Member present smaller diagrams. Within
the units in which higher values for cations such as K+, Na+, and Mg2+,
such as granite, colluvium, and the mix between Pedro Leopoldo
Member and Belo Horizonte Complex, the diagram is partially
asymmetric. The different graphical representations of the waters from
Pedro Leopoldo Member expose the faciological variability of the
limestone of this unit. Like the springs, the waters collected from the
wells are like the ones from the Lagoa Santa Member. On the other hand,
the waters from rocks with frequent pelitic intercalations are hydro­
chemically more distant from the Lagoa Santa Member’s waters due to
smaller concentrations of Ca2+ and HCO−3 (Vieira et al., 2020).

12
F. Aragão et al.
Fig. 10 exposes the hydrochemical compartmentation of the aquifers
proposed by Vieira (2018). Therefore, the evidence suggests that the
lateral compartmentalization result from the lithological changes of
aquifers instead of the chemical evolution, which is also notable within
the subdivisions of Pedro Leopoldo Member’s waters. The classic strat­
igraphic position of the lithological units in the project area is also re­
flected by the vertical variation of the waters in their Si4+
concentrations, which increase in-depth and occur with the content of
silicate minerals in stratigraphically overlapping lithologies. Thus, the
highest concentrations of this ion are verified in the smallest average
quotas of water inputs, a relation notoriously demonstrated by the
hydrochemical types of the Pedro Leopoldo Member. It is mostly marked
by Si+4 concentrations and distribution of the Pedro Leopoldo Member
hydrochemical types. The ones with intermediate and high concentra­
tions of Si+4 (types ii and iii) are situated in the east and southeast of the
study area, which correspond to the more siliciclastic groups of the
Pedro Leopoldo Member and granitic basement. Moreover, the map also
shows the Stiff diagrams of each aquifer unit, including water mixtures,
where it is possible to identify waters with high Ca2+ and HCO3- con­
centrations, corresponding the pure carbonate lithology. Furthermore, it
also indicates waters with higher Mg+2, K+, and Na+, linked to car­
bonate rocks with more pelitic intercalations, granitic rocks, and pelitic
rocks (Vieira, 2018).
Since this study carried a single 7-month sampling campaign in a
relatively large area of 505 km2, there might be limitations for the
presented results. Therefore, it is recommended to implement a solid
monitoring system through periodical sampling campaigns and analyses
to generate consistent water quality data of the region.
6. Conclusions
This paper aimed to propose natural background levels in the region
of The Lagoa Santa Karst Protection Unit for groundwater and springs
and to indicate geogenic anomalies and support the hyrogeochemical
distinction of the aquifer units. This study and its method can be applied
in other karst regions if their inherent characteristics, such as lithology,
hydrogeology, hydrogeochemistry, pedology, hydrography, and land
use inputs, are considered.
The results showed an agreement between the surpassed NBL values
and regional hydrochemistry under the geology’s influence. Among all
the studied species, the geogenic influence is evident, except for chlo­
ride, a possible urban origin. Calcium and bicarbonate are influenced by
the dissolution of carbonates in the Lagoa Santa Member. Magnesium
and silica hold their origins in the pelitic intercalations of the Pedro
Leopoldo Member, which is associated with dolomites. Moreover, high
levels of sodium and potassium are also related to impurities of this
member and pelites present in the Serra de Santa Helena Formation.
Finally, sulfate concentrations are linked to the oxidation of sulfide,
present in rocks both from Lagoa Santa and Pedro Leopoldo members.
Therefore, the integrated analysis between natural background
levels and regional hydrogeochemistry is an important way to indicate
whether high levels of an element or substance are associated with
geogenic or anthropogenic origins. Finally, this method contributes
substantially to the development of studies in karst regions, which are
quite absent in Brazil, targeting mainly the protection of groundwater
resources and water quality.
Declaration of competing interest
The authors declare that they have no known competing financial
interests or personal relationships that could have appeared to influence
the work reported in this paper.
Acknowledgments
Special thanks go to the Foundation for Supporting Research of
13
Journal of South American Earth Sciences xxx (xxxx) xxx
Minas Gerais (FAPEMIG - project APQ-02049-14); to the Post-
graduation course of the Geosciences Institute of the Federal Univer­
sity of Minas Gerais (IGC/UFMG), Brazil; to the Chico Mendes Institute
for Biodiversity Conservation/National Center for Research and Con­
servation of Caves (ICMBIO/CECAV), and the Coordination of Superior
Level Staff Improvement (CAPES).
References
Agenzia per la Protezione dell’Ambiente e per i Servizi Tecnici, Istituto Superiore di
Sanità (APAT-ISS), 2006. Protocollo operativo per la determinazione dei valori di
fondo di metalli e metalloidi nei suoli dei siti contaminati. Giugno, APAT-ISS.
Ajonye, S.E., Bello, I.E., Shaba, H., Asmau, I., Khalid, S., 2016. Remote sensing
assessment of Jabi Lake and its environs: a developmental perspective. Archives of
Agriculture and Environmental Science 1 (1), 1–8, 2456–6632.
ANA – Agência Nacional das Águas, 2012. Conjuntura Dos Recursos Hídricos No Brasil
[Situation of Water Resources in Brazil]. ANA. Brasília. Informe 2012.
Armandine, L.L.A., Aquilina, L., Davy, P., Vergnaud, V., le Carlier, C., 2014. Time scales
of regional circulation of saline fluids in continental aquifers (Armorican massif,
Western France). Hydrol. Earth Syst. Sci. 11, 6599–6635.
Aragão, F., Velásquez, L.N.M., Galvão, P., Tayer, C.C., Lucon, T.N., Azevedo, U.R., 2020.
Natural background levels and validation of the assessment of intrinsic vulnerability
to the contamination in the Carste Lagoa Santa protection unit, Minas Gerais, Brazil.
Environmental Earth Sciences 79, 31. https://doi.org/10.1007/s12665-019-8771-5.
Auler, A., Rubbioli, E., Brandi, R., 2001. As grandes cavernas do Brasil. Grupo Bambuí de
Pesquisas Espeleológicas. Belo Horizonte, 228 p.
Bakalowicz, M., 2005. Karst groundwater: a challenge for new resources. Hydrogeol. J.
13 (1), 148–160.
BRASIL, 2017. Ministério da Saúde. Portaria de Consolidação n 5, de 28 de setembro de
2017. Consolidação das normas sobre as ações e os serviços de saúde do Sistema
Único de Saúde. (Health Minister. Consolidation Ordinance n25, September 28,
2017. Consolidation of the health regulations and services of the Brazil’s Unified
Public Health System). http://bvsms.saude.gov.br/bvs/saudelegis/gm/2017/prc000
5_03_10_2017.html. (Accessed 19 September 2018).
BRIDGE, 2006. Background Criteria for the identification of groundwater thresholds.
Research for policy support. http://nfp-at.eionet.europa.eu/Public/irc/eionet-circle
/bridge/library?l=/public_information/newsletterbridgemay2006p/_EN_1.0_&a=d.
(Accessed 17 November 2018).
BRIDGE, 2009. Background Criteria for the identification of groundwater thresholds.
Research for policy support. http://nfp-at.eionet.europa.eu/Public/irc/eionet-circle
/bridge/library?l=/public_information/newsletterbridgemay2006p/_EN_1.0_&a=d.
(Accessed 17 November 2018).
Bucher, K., Stober, I., 2010. Fluids in the upper continental crust. Geofluids 10, 241–253.
Bucher, K., Zhu, Y., Stober, I., 2008. Groundwater in fractured crystalline rocks, the Clara
mine, Black Forest (Germany). Int. J. Earth Sci. 98, 1727–1739.
CETESB – Companhia Ambiental do Estado de São Paulo, 2011. Guia Nacional de Coleta
e Preservação de Amostras: água, sedimento, comunidades aquáticas e efluentes
líquidos. ANA, Brasília, 326 p.
Chen, Z., Auler, A.S., Bakalowicz, M., Drew, D., Griger, F., Hartmann, J., Jian, G.,
Moosdorf, N., Richts, A., Stevnovic, Z., Veni, G., Goldscheider, N., 2017. The World
Karst Aquifer Mapping project: concept, mapping procedure and map of Europe.
Hydrogeol. J. 25, 771. https://doi.org/10.1007/s10040-016-1519-3.
Cook, P., Love, A., Robinson, N., Simmons, C., 2005. Groundwater ages in fractured rock
aquifers. J. Hydrol. 308, 284–301.
Correa, M.C.P., 2018. The karst under surface analyses: fragility, anthropism and urban
and environmental planning in the Lagoa Santa, MG, karstic region. PhD Thesis.
Federal University of Minas Gerais.
Doerfliger, N., Jeannin, P.Y., Zwahlen, F., 1999. Water vulnerability assessment in karst
environments: a new method of defining protection areas using a multi-attribute
approach and GIS tools (EPIK method). Environ. Geol. 39 (2), 165–176.
Edmunds, W.M., Shand, P., 2008. Natural Groundwater Quality. Blackwell Publishing.
https://doi.org/10.1002/9781444300345.
EPA - United States Environmental Protection Agency, 2000. Guidance for Data Quality
Assessment. Practical Methods for Data Analysis - QA/G-9. QA00. Update. http
s://www.epa.gov/sites/production/files/2015-06/documents/g9-final.pdf.
(Accessed 23 May 2019).
EPA - United States Environmental Protection Agency, 2006. Method 8260C – Volatile
Organic Compounds by Gas Chromatography/Mass Spectrometry. GC/MS. http
s://www.epa.gov/hw-sw846/validated-test-method-8260c-volatile-organic-compou
nds-gas-chromatographymass-spectrometry. (Accessed 15 July 2019).
EPA - United States Environmental Protection Agency, 2014. Method 8270D –
semivolatile organic compounds by gas chromatography/mass Spectrometry. https
://archive.epa.gov/epa/sites/production/files/2015-12/documents/8270d.pdf.
(Accessed 15 July 2019).
European Commission, 1995. Hydrogeological Aspects of Groundwater Protection in
Karstic Areas. Report EUR 16547 EN. Brussels, 446 p.
Ford, D.C., Williams, P.W. (Eds.), 2007. Karst Geomorphology and Hydrology, vol. 2.
Wiley, Chichester, NY, 562 p.
Forti, P., 2002. Speleology in the third millennium: achievements and challenges.
Theoretical and Applied Kartology 15, 7–26.
Foster, S., Hirata, R., Gomes, D., D’Elia, M., Paris, M., 2006. Proteção da qualidade da
água subterrânea: um guia para empresas de abastecimento de água, órgãos
F. Aragão et al.
municipais e agências ambientais. Groundwater Management Advisory Team (GW
Mate). Banco Mundial, Washington, D.C, 104 p.
Galvão, P.H.F., 2015. Hydrogeological Conceptual Model of Sete Lagoas (MG) and
Associated Implications of Urban Development in Karst Region. PhD Thesis. São
Paulo University.
Galvão, P., Hirata, R., Cordeiro, A., Osório, D.B., Peñaranda, J., 2016. Geologic
conceptual model of the municipality of Sete Lagoas (MG, Brazil) and the
surroundings. An Acad. Bras Ciências 88 (1), 35–53.
Galvão, P., Hirata, R., Halihan, T., Terada, R., 2017. Recharge sources and
hydrochemical evolution of an urban karst aquifer, Sete Lagoas, MG, Brazil.
Environmental Earth Sciences 76, 159.
Golden Software, 2017. A basic understanding of surfer gridding methods – Part 1.
https://support.goldensoftware.com/h
c/en-us/articles/231348728-A-Basic-Understanding-of-Surfer-Gridding-Meth
ods-Part-1#. (Accessed 10 August 2018).
Goldscheider, N., Drew, D., 2007. Methods in Karst Hydrogeology: International
Contributions to Hydrogeology, vol. 26. CRC Press, Boca Raton, ISBN 978-0-415-
42873-6.
Gunarathna, M.H.J.P., Kumari, N., Nirmanee, K.G.S., 2016. Evaluation of interpolation
methods for mapping pH of groundwater. IJLTEMAS V (III). ISSN 2278-2540. http
s://www.researchgate.net/publication/299562961/download. (Accessed 30 August
2018).
Hawkes, H.E., Webb, J.S. (Eds.), 1962. Geochemistry in Mineral Exploration. Joanna
Cotler Books, New York, 415 p.
Hinsby, K., Rasmussen, E.S., 2008. The Miocene sand aquifers, Jutland, Denmark. In:
Edmunds, W.M., Shand, P. (Eds.), Natural Groundwater Quality. Blackwell, Oxford,
pp. 306–334.
ISPRA - Istituto Superiore per la Protezione e la Ricerca Ambientale, 2009. Protocollo per
la Definizione dei Valori di Fondo per le Sostanze Inorganiche nella Acque
Sotterranee. https://www.isprambiente.gov.it/files/temi/fondo-metalli-acque-sott
erranee.pdf. (Accessed 10 July 2017).
Jones, W.K., 1977. Karst Hydrology Atlas of West Virginia. Special Publication 4, Karst
Waters Institute, Charles Town, West Virginia, 111 p.
Lucon, T.N., Costa, A.T., Galvão, P., Leite, M.G.P., 2018. Natural background levels and
seasonal influence on groundwater chemistry of the upper São Francisco karst
region, MG, Brazil. Braz. J. Genet. 48 (4), 867–879. https://doi.org/10.1590/2317-
4889201820180071.
Lucon, T.N., Costa, A.T., Galvão, P., 2020. Cadmium behavior in a karst environment
hydrological cycle. Environ. Sci. Pollut. Res. 27, 8965–8979. https://doi.org/
10.1007/s11356-020-07894-2.
Minitab, 2018. Identifying the distribution of your data. https://blog.minitab.
com/blog/meredith-griffith/identifying-the-distribution-of-your-data> Accessed.
(Accessed 28 August 2018).
Mourão, M., Cruz, W., Gonçalves, R., 2001. Caracterização hidrogeológica da porção
mineira da Bacia Hidrográfica do São Francisco. (Hydrogeological characterization
of the mineira portion of São Francisco watershed). In: Bacia do São Francisco:
geologia e recursos naturais. SGB-MG, Belo Horizonte, pp. 327–349.
Mudarra, M., Andreo, B., 2011. Relative importance of the saturated and the unsaturated
zones in the hydrogeological functioning of karst aquifers: the case of Alta Cadena
(Southern Spain). J. Hydrol. 397 (3–4), 263–280. https://doi.org/10.1016/j.
jhydrol.2010.12.005.
National Research Council, 2001. Classifying Drinking Water Contaminants for
Regulatory Consideration. The National Academies Press, Washington, DC. https://
14
Journal of South American Earth Sciences xxx (xxxx) xxx
doi.org/10.17226/10080. https://water.usgs.gov/owq/FieldManual/chapte
r4/pdf/Chap4_v2.pdf.
Paula, R.S., 2019. Modelo conceitual de fluxo dos aquíferos pelíticos-carbonáticos da
Região da APA Carste de Lagoa Santa, MG. PhD Thesis. Federal University of Minas
Gerais.
Persson, A.S., Olsson, O., Rundlöf, M., Smith, H.G., 2010. Land use intensity and
landscape complexity – analysis of landscape characteristics in an agricultural region
in Southern Sweden. Agric. Ecosyst. Environ. 136, 169–176. https://doi.org/
10.1016/j.agee.2009.12.018.
Pessoa, P.F.P., 2005. Hidrogeologia dos aquíferos cársticos cobertos de Lagoa Santa, MG
(Hydrogeology of the Lagoa Santa, MG, covered karstic aquifers). PhD thesis. Federal
University of Minas Gerais. http://www.bibliotecadigital.ufmg.br/dspace/handl
e/1843/ENGD-6LBJ3W. (Accessed 7 June 2017).
Pereira, D.L., Galvão, P., Lucon, T.N., Fujaco, M.A., 2019. Adapting the EPIK method to
Brazilian hydro (geo) logical context of the São Miguel watershed to assess karstic
aquifer vulnerability to contamination. J. S. Am. Earth Sci. 90, 191–203. https://doi.
org/10.1016/j.jsames.2018.12.011.
Reimann, C., Filzmoser, P., Garret, R.G., 2005. Background and threshold comparison of
methods of determination. Sci. Total Environ. 346 (1–3), 1–16. https://doi.org/
10.1016/j.scitotenv.2004.11.023.
Ribeiro, J.H., Tuller, M.P., Danderfer, Filho A., 2003. Geological mapping of the region
of Sete Lagoas, Pedro Leopoldo, Matozinhos, Lagoa Santa, Vespasiano, Capim
Branco, Prudente de Morais, Confins and Funilândia, Minas Gerais State, Brazil
(scale 1:50,000), second ed. Belo Horizonte. 54 p.
Tayer, T.C., 2016. Avaliação da Vulnerabilidade Intrínseca do Aqüífero Cárstico da APA
de Lagoa Santa MG, utilizando o Método COP. Intrinsic Vulnerability Assessment of
Lagoa Santa Protection Unit Karst Aquifer, using COP Method. MS Thesis. Federal
University of Minas Gerais. http://www.bibliotecadigital.ufmg.br/dspace/handl
e/1843/IGCC-AHENHV. (Accessed 8 September 2018).
Tayer, T.C., Velásquez, L.N.M., 2017. Assessment of intrinsic vulnerability to the
contamination of karst aquifer using the COP method in the Carste Lagoa Santa
Environmental Protection Unit, Brazil. Environmental Earth Sciences 76 (13), 445.
https://doi.org/10.1007/s12665-017-6760-0.
Tedd, K., Coxon, C., Misstear, B., Daly, D., Craig, M., Mannix, A., Williams, T.H. (Eds.),
2017. Assessing and Developing Natural Background Levels for Chemical Parameters
in Irish Groundwater. Wexford, Environmental Protection Agency, p. 183.
Tuller, M.P., Ribeiro, J.H., Signorelli, N., Féboli, W.L., Pinho, J.M.M., 2010. Sete Lagoas-
Abaeté Project, Minas Gerais State, Brazil. 6 Geological Maps, Scale 1:100.000
(Geology Program of Brazil), 160 p.
USGS - United States Geological Survey, 2006. National field manual for the collection of
water-quality data. Chapter A4 - collection of water samples. https://water.usgs.
gov/owq/FieldManual/chapter4/pdf/Chap4_v2.pdf. (Accessed 28 August 2018).
Vieira, L.C.M., 2018. Hidrogeoquímica dos Aquíferos da Região da APA Carste de Lagoa
Santa, MG. MS Thesis. Federal University of Minas Gerais.
Vieira, L.C.M., Dussin, T.M., Velásquez, L.N.M., 2018. Geoquímica e condições
paleoambientais de deposição das rochas carbonáticas da Formação Sete Lagoas na
região da Área de Proteção Ambiental Carste de Lagoa Santa, MG. Geochim. Bras. 31
(1), 1–19. https://doi.org/10.21715/GB2358-2812.2018321001.
Vieira, L.C.M., Velásquez, L.N.M., Borba, R.P., 2020. Hidrogeoquímica como ferramenta
para o reconhecimento de aquíferos em lacunas de perfis litológicos de poços
tubulares. Revista Águas Subterrâneas. https://doi.org/10.14295/ras.v34i2.29764.
White, W.B., 1988. Geomorphology and Hydrology of Karst Terrains. Oxford Univ. Press,
New York, 464 p.