Environmental Research 154 (2017) 380–388

Contents lists available at ScienceDirect

Environmental Research
journal homepage: www.elsevier.com/locate/envres

Assessment of heavy metal pollution, spatial distribution and origin in MARK

agricultural soils along the Sinú River Basin, Colombia
⁎ ⁎
José Marrugo-Negretea, , José Pinedo-Hernándeza, Sergi Díezb,
a
University of Córdoba, Faculty of Basic Sciences, Department of Chemistry, Water, Applied and Environmental Chemistry Group, Laboratory of Toxicology
and Environmental Management, Montería, Colombia
b
Department of Environmental Chemistry, Institute of Environmental Assessment and Water Research, IDÆA-CSIC, E-08034 Barcelona, Spain

A R T I C L E I N F O A BS T RAC T

Keywords: The presence of metals in agricultural soils from anthropogenic activities such as mining and agricultural use of
Agricultural soils metals and metal-containing compounds is a potential threat for human health through the food chain. In this
Heavy metal study, the concentration of heavy metals in 83 agricultural soils irrigated by the Sinú River, in northern
Enrichment factor Colombia, affected by mining areas upstream and inundated during seasonal floods events were determined to
Geoaccumulation index
evaluate their sources and levels of pollution. The average concentrations of Cu, Ni, Pb, Cd, Hg and Zn were
Risk assessment code
1149, 661, 0.071, 0.040, 0.159 and 1365 mg/kg respectively and exceeded the world normal averages, with the
exception of Pb and Cd. Moreover, all values surpassed the background levels of soils in the same region. Soil
pollution assessment was carried out using contamination factor (CF), enrichment factor (EF), geoaccumulation
index (Igeo) and a risk assessment code (RAC). According to these indexes, the soils show a high degree of
pollution of Ni and a moderate to high contamination of Zn and Cu; whereas, Pb, Cd and Hg present moderate
pollution. However, based on the RAC index, a low environmental risk is found for all the analysed heavy
metals. Multivariate statistical analyses, principal component and cluster analyses, suggest that soil contam-
ination was mainly derived from agricultural practices, except for Hg, which was caused probably by
atmospheric and river flow transport from upstream gold mining. Finally, high concentrations of Ni indicate
a mixed pollution source from agricultural and ferronickel mining activities.

1. Introduction north for 415 km, mostly through the Córdoba Department, until
flowing into the Caribbean Sea. The soils of the Sinú Valley are
Metal content in soils is the result of metals arising from human primarily used for livestock and agriculture, and the sources of heavy
activities and natural processes. Inputs of metals from anthropogenic metals are mainly influenced by parent materials, mining, application
origins to the soil are much greater than the contribution of metals of agrochemical fertilizer and pesticides, and atmospheric deposition.
from natural sources (Desaules, 2012; Li et al., 2012; Teng et al., Agricultural soils are mainly used for maize (548 km2), cotton
2014). In agricultural soils, the presence of metals is of increasing (227 km2) and rice (96 km2) but also for yucca, kidney beans, red
concern because they have the potential to be accumulated in less pepper, sorghum, bananas and corn. Although regional surveys con-
soluble forms, transferred into soil solution and subsequently deterio- tinue to monitor pesticide residues in water and sediments (Marrugo,
rate groundwater and crop quality (Kelepertzis, 2014). Although 2005; Marrugo et al., 2014, studies on heavy metals in agricultural soils
fertilisers are essential for providing adequate nutrients and ensuring and degrees of pollution are not developed. Therefore, an intensive
successful harvests, long-term repeated application of fertilisers and survey was conducted in the present study to profile the accumulation
metal-containing pesticides and fungicides can gradually accumulate to and sources of Cu, Pb, Zn, Ni, Cd and Hg in 83 soil plots within the
potentially harmful levels in soils (Jiao et al., 2012). Additionally, agricultural system of the Sinú River Basin, covering an area of
because fertilisers tend to be local and ingredient specific, the chemical approximately 300 km2. Several indexes, (i.e., enrichment factor,
composition of soils receiving fertiliser inputs is expected to vary contamination factor and geoaccumulation index) and multivariate
according to application rates and raw mineral sources (Nziguheba and analyses were combined to investigate the current state of soil in this
Smolders, 2008). area and to identify the spatial patterns and possible sources of the
The Sinú River is in northwestern Colombia and flows south to heavy metals. Such research will be used to provide baseline informa-

⁎
Corresponding authors.
E-mail address: [email protected] (J. Marrugo-Negrete) and [email protected] (S. Díez).

http://dx.doi.org/10.1016/j.envres.2017.01.021
Received 3 November 2016; Received in revised form 13 January 2017; Accepted 17 January 2017
Available online 08 February 2017
0013-9351/ © 2017 Elsevier Inc. All rights reserved.
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

tion for the soil quality status in Colombian agricultural soils, as a basis ignition of 2.0g of sediment in an oven at 450 °C for 4 h (Coquery and
for effectively targeting policies to protect soils from long-term Welbourn, 1995). The pH was determined in a relationship soil-water
accumulation of heavy metals and for ensuring food crop quality and 1:1 with a combination glass electrode pH meter coupled to a computer
protecting human health. (WTW, 330i). The soil texture was determined by measuring the
proportions of clay, silt and sand particles present in the soil. These
2. Materials and methods constituents were measured using the pipette method and the soil type
was classified using the soil texture triangle (Pla, 1983). The cation
2.1. Study area exchange capacity (CEC) of the mineral soils was calculated as the sum
of Ca+Mg+K+Na+Fe+Al+Mn extractable with 1 M of NH4-acetate
The Sinú River Basin is located in northwest Colombia, South (IGAC 1979). A sequential extraction scheme was used to study the
America, within the departments of Córdoba and Antioquia in the fractionation of metals in the soils and determine the bioavailability
southwestern region of the Colombian Caribbean coast. The Sinú River and speciation by employing the Bureau of Community Research (BCR)
Basin originates in the Paramillo Massif located between 7° 8′ 9″–9° method (Rauret et al., 2000). Therefore, water-soluble and exchange-
27′ 2″ N and 75° 55′ 31″–75° 58′ 18″ W. The Sinú River is the primary able metals (F1), reducible metals bound to Fe and Mn (F2), oxidisable
water system in the department of Córdoba and its valley is home to metals bound to organic matter and sulphides (F3), and residual metals
agricultural and livestock activities (Feria et al., 2010). Due to the bound to silicate minerals (F4) were investigated. To 1.0g of soil
annual flooding of the Sinú River, the Sinú Valley has some of the most sample, 40 mL of 0.11 mol L-1 acetic acid was added and shaken for
fertile soil in Colombia (Marrugo et al., 2014). Rainfall in this area is 16 h at 25 °C in a mechanical shaker to extract water-soluble,
highly seasonal, with a dry period between January and March, a rainy exchangeable and carbonate-bound metals. The samples were centri-
season between April and December and a short dry period around fuged at 4000 rpm for 10 min and the supernatant passed through
June. An average of 1100–2000 mm of rain falls in this region per year, 0.45 mm filter. To the residue, 40 mL of 0.5 mol L-1 hydroxylamine
with the southern parts of the valley being wetter than those in the hydrochloride was added and shaken for 16 h to extract the reducible
north. The average annual temperature is between 26 and 30 °C (Ruiz metals. To the residue, 10 mL of 8.8 mol L-1 hydrogen peroxide was
et al., 2008). The valley of the Middle Sinú covers an area of 5178 km2 added and digested for an hour at 25 °C and left for another hour at
and contains the municipalities of Monteria, Cereté, San Pelayo, 85 °C in a water bath. The solvent was evaporated and 50 mL of
Ciénaga de Oro and San Carlos. Most of this subregion is composed 1 mol L-1 ammonium acetate was added to the residue and shaken for
of alluvial plains derived from the dynamics of the Sinú River. In 16 h to extract oxidisable metals at pH 2. Finally, aqua-regia digestion
general, the soils are deep or very deep with moderate to high fertility was carried out in a microwave digester to determine the status of the
and are suitable for temporary or permanent crops such as cotton, rice, metals in the residue by atomic absorption analytical method.
maize, beans and pasture (Marrugo et al., 2014). The Lower Sinú valley
covers an area of 1752 km2 and contains the municipalities of Cotorra, 2.3. Quantification of soil pollution
Chimá, Momil, Lorica and Purisima. Soils are composed of recent
alluvial deposits with moderate fertility and are used for agriculture 2.3.1. Contamination factor
and livestock. In this region, land use is divided as follows: 70% is used The contamination factor (CF) is the ratio obtained by dividing the
for livestock, 23% comprising extensive swamps and marshes, and the concentration of each metal in the soil by the baseline or background
remaining 7% is dedicated to agriculture (Viloria de la Hoz, 2004). value:
Cheavy metal
2.2. Soil sampling and chemical analysis CF=
Cbackground
A total of 83 topsoil samples (0–20 cm) were collected from
The contamination levels may be classified based on their inten-
agricultural and pasture areas from the Middle and Lower Sinú River
sities on a scale ranging from 1 to 6 (0=none, 1= none to medium,
Basin during January-March 2015 (dry season). The locations of the
2=moderate, 3=moderate to strong, 4=strongly polluted, 5=strong to
sampling sites are shown in Fig. 1. Each sample was a composite of 5
very strong, 6=very strong) (Muller, 1969). The highest number
subsamples collected from a 10 m×10 m grid using a stainless steel
indicates that the metal concentration is 100 times greater than what
hand trowel and transported to the laboratory. Soil samples were then
would be expected in the crust.
air-dried, passed through a 2 mm polyethylene sieve and finally ground
into fine powder with an agate mortar. The ground samples were stored
in polyethylene zip-type bags for further analysis. 2.3.2. Enrichment factor
For Cu, Zn, Ni, Cd and Pb analysis, the soil samples (0.5 g) were The enrichment factor (EF) for each metal is calculated by dividing
digested with HNO3/HCl 8:2 v/v in a microwave oven (in triplicate) its ratio to the normalising element by the same ratio found in the
using Method 3051 A (USEPA, 2007a). The Hg concentrations were chosen baseline (Taylor, 1964):
determined following the US EPA Method 7471B (USEPA, 2007b), (metal/Al)sample
whereby 0.5 g of soil was digested with H2SO4/HNO3 7:3 v/v and 5% EF=
(metal/Al)background
w/v KMnO4 at 100 °C for 1 h. All analyses were performed using a
spectrometer Thermo Elemental Solaar S4 coupled cold-vapour (Hg), The EF values close to unity indicate crustal origin and EFs > 10
flame (Cu, Zn, Ni), and graphite furnace (Cd, Pb). are considered to be non-crustal source. EF > 1.0 suggest an anthro-
The method for heavy metal estimation was validated with the pogenic origin and those < 1.0 suggest a possible mobilisation or
certified reference material ‘Trace Metals Soil – Sediment’ (CRM008- depletion of metals (Zsefer et al., 1996). The degree of metal pollution
050: certified value; Hg 0.72 mg/kg, Cu 36.4 mg/kg, Ni 26 mg/kg, Pb was based on seven enrichment factor classes (Taylor, 1964).
95.3 mg/kg, Cd 0.82 mg/kg). The percentages of recovery average for Accordingly, an EF value of 50 corresponds to extremely severe
heavy metals were 98.4% (n=3). The sequential extraction method enrichment; EF=25 – < 50 to very severe enrichment; EF=10 – < 25
showed no statistically significant differences (P < 0.05) between the ∑ to severe enrichment; EF=5 – < 10 to moderately severe enrichment;
Heavy metal of the extracted fractions and the concentrations of heavy EF=3 – < 5 to moderate enrichment; EF=1 – < 3 to minor enrichment
metals determined by digestion of the soils, demonstrating that the and EF < 1 to no enrichment. For most heavy metals of environmental
method shows good precision ( < 2.5%) and accuracy ( < 5.0%). interest, concentrations in soil easily vary over two to three orders of
The organic matter was determined as the percentage loss on magnitude depending on the parent materials (Blaser et al., 2000). In

381
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

Fig. 1. Location of the Lower and Middle Sinú River basin and the sampling stations of the agricultural soils collected in each part of the basin.

this regard, the background values in the present study (Table 1) were Cn
lgeo = log2
calculated from the mean concentrations of heavy metals analysed in 1.5Bn
soils with no agricultural management from unflooded zones in the
Lower and Middle Sinú River. where Cn is the concentration of metals examined in soil samples and
Bn is the geochemical background concentration of the metal (n).
Factor 1.5 is the background matrix correction factor due to lithologic
variation in the soils (Stoffers et al., 1986). The lgeo consists of seven
2.3.3. Geoaccumulation index grades or classes (Muller, 1981): Class 0 (uncontaminated) Igeo ≤0;
Geoaccumulation index (Igeo) for the metals were determined Class 1 (from uncontaminated to moderately contaminated) 0 < Igeo
using (Muller, 1979): < 1; Class 2 (moderately contaminated) 0 < Igeo < 2; Class 3 (from

382
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

Table 1
Statistical summary of chemical elements in soils (n=83) from Sinú River basin (in mg/kg).

Sinú River basin Cu Ni Pb Cd Hg Zn

Low Sinú (n=45)

All crops Mean 1004 587 0.066 0.035 0.177 1218

Median 1171 592 0.065 0.033 0.157 1112
SD 560 384 0.020 0.015 0.080 547
Minimum 12.6 8.3 0.02 0.010 0.050 285
Maximum 1908 1757 0.10 0.060 0.410 2233
Rice (mean ± SD) 754 (692) 489 (338) 0.058 (0.014) 0.029 (0.013) 0.207 (0.098) 955 (455)
Pasture (mean ± SD) 1155 (420) 591 (327) 0.072 (0.021) 0.040 (0.014) 0.161 (0.062) 1422 (521)
Cotton (mean ± SD) 636 (425) 890 (640) 0.059 (0.023) 0.031 (0.025) 0.116 (0.052) 961 (605)

Middle Sinú (n=38)

All crops Mean 1321 749 0.077 0.045 0.139 1540

Median 1251 787 0.082 0.049 0.128 1642
SD 438 289 0.021 0.012 0.079 424
Minimum 490 130 0.04 0.02 0.04 769
Maximum 2522 1391 0.13 0.08 0.37 2632
Rice (mean ± SD) 926 (296) 479 (266) 0.059 (0.024) 0.035 (0.014) 0.093 (0.040) 1125 (437)
Pasture (mean ± SD) 1240 (354) 671 (270) 0.066 (0.028) 0.039 (0.016) 0.170 (0.090) 1329 (553)
Cotton (mean ± SD) 1426 (424) 835 (254) 0.083 (0.013) 0.049 (0.008) 0.137 (0.089) 1649 (255)

Low and Middle Sinú (n=83)

All crops Mean 1149 661 0.071 0.040 0.159 1365

Median 1213 681 0.073 0.043 0.149 1497
SD 529 351 0.022 0.015 0.081 518
Minimum 12.6 8.3 0.02 0.01 0.04 285
Maximum 2522 1757 0.13 0.08 0.41 2632
Crustala Mean 55 20 15 0.1 0.07 70
World-soilb Mean 38.9 29 27 0.41 0.07 70
Unpolluted Colombian soil (n=5) Mean (SD) 118.1 (2.4) 14.1 (0.8) 0.012 (0.001) 0.008 (0.001) 0.028 (0.007) 107 (1.8)

a
Mean values compiled from Mason and Moore (vide Hedrick, 1995) and Reimann and de Caritat (1998).
b
Mean values compiled from Kabata-Pendias (2011)

moderately to strongly contaminated) 2 < Igeo < 3; Class 4 (strongly contamination in the study area and source identification, PCA was
contaminated) 3 < Igeo < 4; Class 5 (from strongly to extremely used with Varimax rotation to minimise the number of variables with a
contaminated) 4 < Igeo < 5; Class 6 (extremely contaminated) 5 < high loading on each component. Hierarchical CA was developed
Igeo. Class 6 is an open class and comprises all values of the index according to the average linkage (between groups) method and the
higher than Class 5. The elemental concentrations in Class 6 may be Pearson's correlation as a similarity measure. The results are reported
hundred-fold greater than the geochemical background value. in the form of a dendrogram, thus providing a visual summary of the
cluster. The results from the PCA were interpreted according to the
2.3.4. Risk assessment code hypothetical sources of chemical elements. Statistical analysis was
The Risk Assessment Code (RAC) considers the ability of metals to performed with SPSS v23.0.0.0.
be released and subsequently enter into the food chain and is based on
the strength of the bond between metals and other components in soil. 3. Results and discussion
Therefore, the RAC assesses the availability of metals by applying a
scale to the percentage of metal in the carbonate and exchangeable 3.1. Heavy metal concentrations of soil samples
fractions (i.e., F1). When the percentage F1 is less than 1% there is no
risk (NR). For a range of 1–10%, there is low risk (LR), medium risk The agricultural soils irrigated by the Sinú River waters and flooded
(MR) for a range of 11–30%, high risk (HR) for 31–50%, and very high during flooding events contain high levels of metal, although their
risk (VHR) for 51–100% (Sundaray et al., 2011). contents vary depending on the location of the soils. The descriptive
statistics of the heavy metal concentrations in agricultural soils of the
2.4. Statistical analysis Sinú Basin are listed in Table 1. The median values of elemental
contents in soils follow a decreasing order as: Zn > Cu > Ni > Hg > Pb >
The results for each sample were calculated as the mean ± the Cd, in both the middle and lower Sinú River Basin and if all sampling
standard deviation from triplicate determinations. A Student's t-test points are considered altogether. Mean levels of all the analysed heavy
was used to evaluate if there were significant differences between the metals in the agricultural soils were higher than the corresponding
average concentrations from two different sampling stations or expedi- values determined for the reference soils (i.e., unpolluted Colombian
tions. The criterion for significance was set at P < 0.05. soil) (Table 1), and except for Pb and Cd, values were higher than
Multivariate cluster analysis (CA) and principal component analysis crustal and average worldwide soils. Moreover, levels for Ni are 50-
(PCA) were employed with the aim of identifying associations and times higher than predicted from unpolluted Colombian soils
common origin among metals. Thus, CA was applied to identify (14.1 mg/kg) and 5-times to 13-times higher for the rest of the metals.
different geochemical groups, clustering the samples with similar heavy Heavy metal concentrations present statistically significant differences
metal contents; whereas, for further evaluation of extent of metal between sampling stations for the middle and lower areas (P < 0.05).

383
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

The highest concentration of most heavy metals was found to be

associated with the middle reaches of the Sinú River, except for Hg,
which is related with the lower part of the river, characterised as a
floodplain (Feria et al., 2010). Poot et al. (2007) indicated that the
phenomenon of flooding soil for a sufficiently long period could provide
adequate reduction of sulphates and formation of volatile acids and the
subsequent release and availability of heavy metals. This could explain
the common source of Cu, Ni and Zn from the lithology of the
floodplain of the lower Sinú. On the other hand, it is well know that
anthropogenic activities such as metal mining and smelting, industrial
activities, irrigation, municipal waste disposal, fertilisation and inten-
sive use of pesticides and herbicides result in heavy metal accumulation
in soils (Smith, 2009; Liu et al., 2011; Teng et al., 2014; Chen et al.,
2015). The use of the phosphate fertilizers will have had some impact
on introducing some of these metals to soils of the Sinú River basin, but
crop type, rate of fertilizer applications, and flooding/irrigation fre-
quency will all have a role in determining the concentrations that
accumulate. Noteworthy was the frequent use of superphosphate and Fig. 2. Distribution of metals in agricultural soils depending on the crop. Error bars are
other phosphate fertilisers in this area (Marrugo, 2005; Lans et al., standard deviations of the average field sample (n=3).
2008), which contain concentrations of Cu, Ni, Pb, Cd and Zn and the
addition of pesticides containing Cu and Zn, as well as herbicides and (0.207 mg/kg) with respect to the Middle Sinú valley (0.093 mg/kg).
fungicides containing Hg. Over 50 years ago, the use of fungicide Furthermore, no differences were found when the other crops were
containing phenylmercury was very common in this area as well as analysed for Hg in both geographical areas. Numerous studies have
massive aerial fumigation of paddy fields with mercurial agrochem- identified rice paddy fields as important sites for methylmercury
icals, which were prohibited by ICA (Colombian Agricultural Institute) production, reporting high levels in rice grain and in the corresponding
in resolution 2189 in 1974. Nevertheless, it should be noted that Hg soil in Hg mining areas and close locations (Zhang et al., 2010; Meng
levels are probably also associated with atmospheric Hg deposition et al., 2014). Moreover, recent studies suggest that bioavailable Hg
(wet and dry), due to the proximity of specific sources of Hg, such as derived from new atmospheric deposition appears to be the primary
gold mining areas located a few km northwest in the Mojana Region factor regulating net methylmercury production in rice paddies (Zhao
(Pinedo-Hernández et al., 2015; Marrugo-Negrete et al., 2015). et al., 2016a, 2016b). These studies propose that the in situ production
Furthermore, the heavy metal content can be associated with natural in paddy water and surface soil is dependent on elevated Hg in the
concentrations of the soil, urbanisation processes in different popula- atmosphere and deposition of new Hg into a low-pH anoxic ecosystem.
tions within the study area, as well as the wastewater discharged by Therefore, we hypothesize that atmospheric Hg deposition and,
urban districts into the Sinú River that via the flood dynamics can especially, frequent flooding increases Hg loads in soils of paddy fields
deposit metals in soils (Marrugo, 2005; Feria et al., 2010) In this mostly located at the Lower Sinú, due to upstream gold mining sites.
regard, long-term use of wastewater irrigation for agricultural produc- Even though our hypothesis is in agreement with the observed facts of
tion is a significant source of soil heavy metal pollution worldwide Zhao et al., data that supports this hypothesis are lacking, and further
(Gupta et al., 2012; Lu et al., 2015). Another possible source of heavy work about the fate of Hg and MeHg production in the soil-rice system
metal enrichment may be the influence of the Urrá dam, due to the in the Sinú River is needed.
flooding of 7400 ha of soil, which accelerated the process of washing The percentage of organic matter (OM) in the analysed soils is
and stripping of metals in addition to constituent elements of the relatively low (mean of 1.86%), with variations from 0.17% to 3.55%,
affected soil by pouring water down the dam (Feria et al., 2010). indicating the bioavailability of heavy metals and presenting low
Previous studies identified the presence of a high concentration of Cu, formation of organometallic complexes based on the OM content
Ni, Cr and Zn in surficial sediments of river Sinú (Feria et al., 2010), (Silveira et al., 2003; Basta, 2005). The pH controls adsorption/
and high Hg levels in carnivorous fish species (Ruiz-Guzman et al., desorption and precipitation/solubilisation reactions (Basta, 2005)
2014) of the Urrá dam that could be linked to the recent opening of and can be considered the most influential factor, causing an increase
many artisanal fronts of gold mining upstream of the reservoir. in solubility and therefore the bioavailability of the elements inversely
Moreover, a recent study revealed that in the upper part of the basin proportional to soil pH (Silveira et al., 2003; Antoniadis et al., 2008).
at the municipality of Tierralta there are important alluvial and hard- In total 84% of the soils present acidic characteristics, which may
rock gold mining activities, that use mercury to recover gold (UPME- favour the bioavailability of heavy metals, whereas, 16% have basic
MME-UC, 2016). features that can promote adsorption of metals, thus decreasing
On the other hand, heavy metals show a similar pattern depending bioavailability (Lasat, 2002). Soil textures have sand content of 32–
on the land use. The trends for Cu, Ni, Zn, Cd and Pb reveal that these 86%, clay 3–49%, silt 4–49% and in the study area sandy clay soils are
elements follow an increasing order in: rice > pasture > cotton; how- predominant. The CEC of soil is moderate (17.6–40.4 meq/100 g),
ever, soils cultivated for rice show higher Hg levels than pasture and with an average of 26.2 meq/100 g. The type of substrate and its
cotton (Fig. 2). Statistical analyses show that there are significant texture are crucial in metal fixation and also depends on the clay
differences (P < 0.05) for Cu, Cd and Zn when paddy fields are content of the soil (Kabata-Pendias, 2011). Based on the results of
compared with soils cultivated with pasture and/or cotton. Moreover, texture and CEC, it can be inferred that the soils in the study area tend
Ni, Pb and Hg do not show significant differences when rice and to retain heavy metal cations because the CEC is moderately related to
pasture soils are compared. Interestingly, no significant differences the clay content that decreases its bioavailability. Studies by Combatt
were found for Hg (P > 0.05) when different agricultural soils were et al. (2003, 2005) reported similar results of pH, OM and soil textural
matched, which suggests that the Hg content in soils is not related to composition in the Sinú Valley, indicating that the characteristics are
the land use or the application of different crop dependent pesticides. maintained as a function of time despite the environmental conditions
Nevertheless, when Hg content of the different crops was compared by to which the study area may be exposed. Values of heavy metals, pH,
different geographical area (e.g., Lower Sinú vs. Middle Sinú), only OM and CEC in sampling sites can be found in supplementary material
paddy fields show significantly higher levels (P=0.026) in the Lower (Table S1).

384
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

Fig. 3. The percentage of heavy metal in different fractions in the agricultural soils.
Mean ± SD. F1: water-soluble and exchangeable; F2: reducible metals; F3: OM-
sulphides; F4: residual metals.

3.2. Bioavailability of heavy metals

Fig. 3 shows the percentage of the chemical fractions of Ni, Hg, Cd,
Cu, Zn found in soils, with the order F3 > F2 > F4 > F1. Metal content
in soils mainly associated with the F3 (OM-sulphides), represents 47%
(Ni), 56% (Hg), 65% (Cd), 46% (Pb), 53% (Zn) and 55% (Cu) of the
total content. This fraction, although being a large proportion, has a
moderate mobility. In this phase, metals may be associated with
organic ligands (e.g., humic and fulvic acids). There are also reduced
sulphur species that bind to metals such as Hg (Xia et al., 1999)
present in the OM. The F2 fraction represents 29% (Ni), 29% (Hg),
24% (Cd), 39% (Pb), 31% (Zn) and 32% (Cu), showing that a great part
of metals at the different sampling stations is associated with amor-
phous aluminosilicate materials, such as oxides of Fe and Mn. The
residual or lithogenic fraction (F4) represents 17% (Ni), 7% (Hg), 6%
(Cd), 12% (Pb), 6% (Zn) and 9% (Cu). In this phase, the metals are
bound to minerals as part of their crystal structures. The release of
metals in this phase within a reasonable period of time is unlikely. The
bioavailable fraction (F1) represents 4% (Ni), 1% (Hg), 3% (Cd), 3%
(Pb), 5% (Zn) and 4% (Cu). Therefore, since the RAC is calculated
based on the percentage of F1, according to our results the environ-
mental risk is low for all the analysed heavy metals.

3.3. Assessment of soil pollution

The CF, Igeo and EF values in soils are summarised in Fig. 4. The
range of CF for the different metals was 0.7–21 (Cu), 0.6–125 (Ni),
1.9–11 (Pb), 1.1–10 (Cd), 1.5–15 (Hg) and 2.7–25 (Zn). According to
our results (Fig. 4a), the CF average for all metals follow the decreasing
order as: Ni (47) > Zn (13) > Cu (10) > Pb (6)=Hg (6) > Cd (5). The CF
average based on background values for all metals indicates contam-
ination by human sources (CF > 1), with metals such as Pb, Cd and Hg
presenting moderate degrees of contamination (CF of 3–6), whereas
Cu, Zn and Ni showed high contamination (CF > 6). In comparison
with other countries, values of CF in different soils collected in the
Lower and Middle Sinú River are higher than those reported along the
Jishui River, China (Liu et al, 2013) and in agricultural soils in
northern Bangladesh (Bhuiyan et al., 2010).
On the other hand, EF values calculated for each one of the metals
studied are shown in Fig. 4b. EF values were interpreted with the
support of the classification of Taylor (1964). The mean EF values
suggest minor enrichment for Ni (EF: 1.79), unlike Zn (0.49), Cu Fig. 4. Box and whisker plots display the distributions of the different contamination
(0.37), Pb (0.23) Hg (0.21), Cd (0.19) with EF values < 1.0, indicating indexes: Contamination Factor (a), Enrichment Factor (b), and Geoaccumulation index
no enrichment. Cd and Pb are anthropogenic metals usually generated (c). Boxes depict 25th, 50th (median), and 75th percentiles and whiskers minimum and
by human activities. However, while no enrichment was showed, these maximum values. Mean values (О); outliers (*).

metals can be associated in a greater extent with colloidal materials in

385
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

Table 2
Pearson correlation matrix for heavy metal in soil samples.

Cu

Cu 1 Ni

Ni 0.327* 1 Pb
Pb 0.620** 0.269* 1 Cd
Cd 0.687** 0.325* 0.953** 1 Hg
Hg −0.050 0.084 0.028 −0.056 1 Zn
Zn 0.718** 0.343* 0.917** 0.965** −0.058 1

*
Correlation is significant at the 0.05 level.
**
Correlation is significant at the 0.01 level.

surface runoff that can be easily transported in the Sinú River flow,
which could reach the fields as a result of flooding. Enrichment of Ni
can be attributed to the extraction process of ferronickel smelting
plants in the San Jorge River Basin, located relatively close to our study
area. Previous studies reported elevated concentrations of Ni in rainfall
and dry deposition at locations 30 km away from the ferronickel
mining area (Marrugo-Negrete et al., 2014).
The Igeo was also used as a reference for estimating the extent of
metal pollution (Fig. 4c). Mean Igeo values are in the order Ni (4.6 ±
1.5) > Zn (3.0 ± 0.7) > Cu (2.4 ± 1.1) > Pb (1.9 ± 0.5) > Hg (1.8 ± 0.8)
> Cd (1.6 ± 0.7). Therefore, the Igeo indicates moderate to heavy
pollution of metals in the study area, depending on each metal and
sampling station location. Then, Pb, Cd and Hg present a moderate
contamination (1 < Igeo < 2), while Zn and Cu present moderate to
heavy contamination (2 < Igeo < 3). Again, Ni is significantly accumu-
lated in the soils and its Igeo indicates heavily pollution (4 < Igeo < 5).
Inter-element relationships in the soil matrix provide information
on heavy metal sources and pathways in the geoenvironment (Dragovic
et al., 2008). According to the values of Pearson correlation coefficients
(Table 2), except for Hg, the rest of the heavy metals are all positively
correlated. A significant strong positive correlation exists between Cd
and Zn (r=0.965, P < 0.01), Cd and Pb (r=0.953, P < 0.01) and Pb and
Fig. 5. Principal Component Analysis (PCA) results. (a) Loading plot of the two
Zn (r=0.917, P < 0.01). Ni shows a relatively weak positive correlation components influencing geochemical variation of agricultural soils from Sinu River
with Cu (r=0.327, P < 0.05), Pb (r=0.269, P < 0.05) and Cd (r=0.325, P basin, (b) Score plot PC1 vs. PC2 of sampling sites. White circles: S1-S45: lower Sinú
< 0.05) and significant moderate positive correlations exist between Cu River: black squares: S46-S83: middle Sinú River. Codes in bold indicate sampling sites
and Pb (r=0.620, P < 0.01) and Cd (r=0.687, P < 0.05). with relatively high scores in first or second component.
These significant and relatively strong correlations indicate that
these metals in the Sinú River soils were derived from similar sources, that shows an increased accumulation in soils after long-term direct
which mainly originate from anthropogenic activities. For further use of Cu-based pesticides and fungicides (Epstein and Bassein, 2001).
evaluation of the extent of metal contamination in the study area and Increased Pb levels in cultivated soils have also been ascribed to
source identification, PCA was used following the standard procedure mineral fertiliser application, mainly phosphate fertilisers (Nziguheba
reported in literature (Dragovic et al., 2008). This technique clusters and Smolders, 2008). Cd accumulation in agricultural soils under
variables into groups, such that variables belonging to one group are intense application of phosphate fertiliser with high Cd content is also
highly correlated with one another. well documented (Atafar et al., 2010). Zn is proposed to exhibit a
The results of the PCA of heavy metal contents and sampling points mixed source of both lithogenic and human origin in agricultural soils
are shown in Fig. 5. Two principal components with eigenvalues (Keleperzis et al., 2014). Animal manure is suggested to contribute to
greater than 1 (before and after rotation) were extracted. The PCA Zn levels in agricultural soils (Nicholson et al., 2003). However, the
method leads to a reduction of the initial dimension of the dataset to high association of Zn with Cd as revealed by the Pearson correlation
two components explaining 77.5% of the data variation. Therefore, (Table 2) and CA (Fig. 6) points out that the application of phosphate
these two factors play a significant role in explaining metal contamina- fertilisers is the most important source of Zn in the agricultural system
tion in the study area. As shown Fig. 5a the rotated component matrix under study here.
demonstrated that Cu, Pb, Cd and Zn were involved in the first An atypical behaviour was revealed in the present study for Ni. The
component (PC1) whereas the second component (PC2) included Hg. moderate loadings of this element observed on PC1 and PC2 suggest
Ni was the only element that did not demonstrate a clear association that this metal probably has anthropogenic sources but that these are
with either the first or the second component and had similar moderate different than those related with agricultural activities. Previous studies
loading plots for PC1 (0.424) compared to PC2 (0.454). In more detail, reveal that Ni concentrations were detected in atmospheric deposition
PC1, which has the highest loadings of Cu, Pb, Cd and Zn accounts for (rain and dry) and could be attributed to the extraction process of
60% of variance and is the most important component. PC1 could be ferronickel smelting plants, located a few kilometres north of our study
better explained as an anthropogenic source, specifically related to the site (Marrugo-Negrete et al., 2014).
agricultural activities taking place in the area for a long period of time. PC2 accounts for 17.5% of the total variability and can be presumed
Cu is a well-known marker element of agricultural activities, to represent anthropogenic sources, however from a different origin
specifically related to application of fertilisers (Acosta et al., 2011) than those related to farming. Atmospheric and river flow transport of

386
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

Hg (dissolved and associated with particulate matter) from upstream

gold mining activities seem to be the main sources of Hg in the
agricultural soils of the Sinú River Basin. We hypothesize that inputs of
Hg occur through dry and wet deposition as well as flooding of fields
during flow events in the rainy season due to the close location of gold
mining areas in the Mojana region (Pinedo-Hernández et al., 2015;
Marrugo-Negrete et al., 2015). Based on information assessed from
PCA, hierarchical CA was performed and the results are illustrated in
Fig. 6. Dendrogram in Fig. 6a enabled the identification of two major
clusters. Cluster 1 is composed of Cu, Pb, Zn, Cd and Ni, in agreement
with PC1. Thus, cluster 1 is probably associated with anthropogenic
sources from agricultural management practices, such as fertilisation
and pesticide application. The second cluster is composed of Hg and
can be presumed to represent an anthropogenic origin different to that
of cluster 1. The obvious enrichment of Hg in agricultural soils of the
Sinú River Basin is ascribed to the presence of upstream gold mining
sites.
Similarly, sampling points were also analysed by clustering meth-
ods (Table S2 and biplot in Fig S2) and organized in the dendrogram to
identify similar groups (Fig. 6b). The sampling sites could be grouped
in two clusters, with the majority in cluster 1 with 62 samples, and
cluster 2 with 21 samples. In addition, among all the sampling sites,
75% represent cluster 1, with a similar percentage of samples located in
the Lower Sinú River (48%) versus the Middle Sinú River (52%).
Interestingly, in cluster 2, a 71% of the sampling sites are located in the
Lower Sinú River. This distribution is also observed in Fig. 5b, where
sites codes in Lower Sinú (white circles) and Middle Sinú (black
squares) with high scores in PC1 or PC2 are marked in bold. There are
12 sampling sites with high scores in PC1 ( > 1): S60, S46, S58, S39,
S59, S14, S56, S57, S72, S40, S73 and S63, which could be grouped in
cluster 1. Among them, a 75% correspond with the Middle Sinú River,
which are enriched with Cu, Pb, Cd, Zn and Ni. The sampling points
with relatively high scores in PC2 ( > 1) are 14 sites: S48, S31, S07, S70,
S09, S34, S20, S74, S35, S50, S49, S21, S10 and S08 are grouped in
cluster 2. Among all these sites, a 90% have negative scores in PC1,
being most of them paddy fields in the Lower Sinú River, which are
loaded with Hg.

4. Conclusions

The present study showed marked variations and diverse correla-

tions of selected metal contents. The distribution of selected metals in
soils exhibited spatial variations, while, multivariate analysis revealed
significant anthropogenic pollution from the selected metals in the
soils. Results of different contamination indexes supported by the
application of factor, cluster and correlation analysis highlight the clear
anthropogenic impact of specific metals (Cu, Zn, Cd, Pb) in agricultural
soils due to the application of fertilisers and metal-containing pesti-
cides and fungicides. Although Cu, Ni, Zn and Hg were present in large
amounts in the studied soils compared to in crustal and average
worldwide soils, environmental risk was found to be low. Multivariate
analysis showed that heavy metal contamination in soils mainly
originates from anthropogenic origin though agricultural activities.
The only metal that was evaluated to exhibit a mixed source was Ni,
which is attributed to both the application of agrochemicals and
ferronickel mining activities. The spatial distribution trends reveal that
the high concentrations of Hg are associated with the paddy fields and
significantly elevated levels are found in the Lower Sinú River, which
Fig. 6. Dendrogram obtained by hierarchical clustering analysis for (a) the heavy metals, are probably more impacted by atmospheric Hg deposition and flood-
and (b) the sampling sites. ing events from upstream gold mining activities. Future studies should
pay attention to characterise the fractionation and reactivity of metals
in agricultural soils by utilising selective chemical extractions with the
aim of assessing the actual risks for the environment.

387
J. Marrugo-Negrete et al. Environmental Research 154 (2017) 380–388

Acknowledgements Marrugo-Negrete, J.L., Navarro-Frómeta, A.E., Urango-Cardenas, I.D., 2014.

Organochlorine pesticides in soils from the Middle and Lower Sinú River Basin
(Córdoba, Colombia). Water. Air. Soil. Pollut. 225, 2053.
We appreciate the anonymous reviewers for their valuable com- Marrugo, J.L., 2005. Determinación de los niveles de pesticidas en los ríos Sinú, San
Jorge y Canalete, con el fín de darle soporte al diagnostico del plan de ordenamiento
ments and suggestions to improve the quality of the manuscript. The Territorial de las cuencas. CVS – Universidad de Córdoba.
authors wish to thank the University of Cordoba, Montería-Colombia. Marrugo-Negrete, J., Pinedo-Hernández, J., Díez, S., 2015. Geochemistry of mercury in
Water, Applied and Environmental Chemistry Group, Laboratory of tropical swamps impacted by gold mining. Chemosphere 134, 44–51.
Marrugo-Negrete, J.L., Urango-Cardenas, I.D., Burgos Núñez, S.M., Díez, S., 2014.
Toxicology and Environmental Management, University of Cordoba, Atmospheric deposition of heavy metals in the mining area of the San Jorge river
Montería-Colombia. basin, Colombia. Air Qual. Atmos. Health 7, 577–588.
Meng, B., Feng, X.B., Qiu, G.L., Anderson, C.W.N., Wang, J.X., Zhao, L., 2014.
Localization and speciation of mercury in brown rice with implications for Pan-Asian
Appendix A. Supplementary material public health. Environ. Sci. Technol. 48, 7974–7981.
Muller, G., 1969. Index of geoaccumulation in sediments of the Rhine River. Geojournal
2, 108–118.
Supplementary data associated with this article can be found in the Muller, G., 1979. Schwermetalle in den sediments des Rheins-Veranderungen seitt 1971.
online version at doi:10.1016/j.envres.2017.01.021. Umschan 79, 778–783.
Muller, G., 1981. Die Schwermetallbelstung der sedimente des Neckars und seiner
Nebenflusse: eine Bestandsaufnahme. Chem. Ztg. 105, 157–164.
References Nicholson, F.A., Smith, S.R., Alloway, B.J., Carlton-Smith, C., Chambers, B.J., 2003. An
inventory of heavy metals inputs to agricultural soils in England and Wales. Sci.
Total Environ. 311, 205–219.
Acosta, J.A., Faz, A., Martínez-Martínez, S., Arocena, J.M., 2011. Enrichment of metals
Nziguheba, G., Smolders, E., 2008. Inputs of trace elements in agricultural soils via
in soils subjected to different land uses in a typical Mediterranean environment
phosphate fertilizers in European countries. Sci. Total Environ. 390, 53–57.
(Murcia city, southeast Spain). Appl. Geochem. 26, 405–414.
Pinedo-Hernández, J., Marrugo-Negrete, J., Díez, S., 2015. Speciation and bioavailability
Antoniadis, V., Robinson, J., Alloway, B., 2008. Effects of short-term pH fluctuations on
of mercury in sediments impacted by gold mining in Colombia. Chemosphere 119,
cadmium, nickel, lead, and zinc availability to ryegrass in a sewage sludge-amended
1289–1295.
field. Chemosphere 71, 759–764.
Pla, I., 1983. Metodología para la caracterización física con fines de diagnóstico de
Atafar, Z., Mesdaghinia, A., Nouri, J., Homaee, M., Yunesian, M., Ahmadimoghaddam,
problemas de manejo y conservación de suelos en condiciones tropicales. Rev. Fac.
M., Mahvi, A.H., 2010. Effect of fertilizer application on soil heavy metal
Agron. (UCV), Alcance 1983 (33), 15–91.
contamination. Environ. Monit. Assess. 160, 83–89.
Poot, A., Gillissen, F., Koelmans, A., 2007. Effects of flow regime and flooding on heavy
Basta, N.T., Ryan, J.A., Chaney, L., 2005. Trace element chemistry in residual-treated
metal availability in sediment and soil of a dynamic river system. Environ. Pollut.
soil: key concepts and metal bioavailability. J. Environ. Qual. 34, 49–63.
148, 779–787.
Bhuiyan, M.A.H., Parvez, L., Islam, M.A., Dampare, S.B., Suzuki, S., 2010. Heavy metal
Rauret, G., López-Sánchez, J.F., Sahuquillo, A., Barahona, E., Lachica, M., Ure, A.M.,
pollution of coal mine-affected agricultural soils in the northern part of Bangladesh.
Davidson, C.M., Gomez, A., Lück, D., Bacon, J., Yli-Halla, M., Muntau, H.,
J. Hazard. Mater. 173, 384–392.
Quevauviller, P., 2000. Application of a modified BCR sequential extraction (three-
Blaser, P., Zimmermann, S., Luster, J., Shotyk, W., 2000. Critical examination of trace
step) procedure for the determination of extractable trace metal contents in a sewage
element enrichments and depletions in soils: As, Cr, Cu, Ni, Pb, and Zn in Swiss
sludge amended soil reference material (CRM 483), complemented by a three-year
forest soils. Sci. Total Environ. 249, 257–280.
stability study of acetic acid and EDTA extractable metal content. J. Environ. Monit.
Chen, H., Teng, Y., Lu, S., Wang, Y., Wang, J., 2015. Contamination features and health
2, 228–233.
risk of soil heavy metals in China. Sci. Total Environ. 512, 143–153.
Reimann, C., de Caritat, P., 1998. Chemical Elements in the Environment. Springer-
Combatt, E., Palencia, G., Marin, N., 2003. Clasificación de suelos sulfatados ácidos
Verlag, Berlin, 398.
según azufre extraíble en los municipios del medio y bajo Sinú en Córdoba. Temas
Ruiz, M., Bernal, G., Polania, J., 2008. Influencia del Río Sinú y el mar Caribe en el
Agrar. 8, 22–29.
sistema lagunar de Cispatá. Bol. Invest. Mar. Cost. 37, 29–49.
Combatt, E., Martínez, G., Polo, J., 2005. Caracterización química y física de los suelos
Ruiz-Guzman, J., Marrugo-Negrete, J.L., Diez, S., 2014. Human exposure to mercury
agroforestales de la zona alta de Córdoba. Temas Agrar. 10, 5–14.
through fish consumption: risk assessment of riverside inhabitants of the Urrá
Coquery, M., Welbourn, P.M., 1995. The relationship between metal concentration and
reservoir. Colomb. Hum. Ecol. Risk Assess. 20, 1151–1163.
organic matter in sediments and metal concentration in the aquatic macrophyte
Silveira, M., Alleoni, L., Guilherme, L., 2003. Biosolids and heavy metals in soils. Sci.
Eriocaulon septangulare. Water Res. 29, 2094–2102.
Agric. 60 (4), 793–806.
Desaules, A., 2012. Critical evaluation of soil contamination assessment methods for
Smith, S.R., 2009. A critical review of the bioavailability and impacts of heavy metals in
trace metals. Sci. Total Environ. 426, 120–131.
municipal solid waste composts compared to sewage sludge. Environ. Int. 35,
Dragovic, S., Mihailovic, N., Gajic, B., 2008. Heavy metals in soils: distribution,
142–156.
relationship with soil characteristics and radionuclides and multivariate assessment
Stoffers, P., Glasby, G.P., Wilson, C.J., Davis, K.R., Watter, P., 1986. Heavy metals
of contamination sources. Chemosphere 72, 491–549.
pollution in Wellington harbour. N.Z. J. Mar. Freshw. Res. 20, 495–512.
Epstein, L., Bassein, S., 2001. Pesticide application of copper on perennial crops in
Sundaray, S.K., Nayak, B.B., Lin, S., Bhatta, D., 2011. Geochemical speciation and risk
California, 1993 to 1998. J. Environ. Qual 22, 335–348.
assessment of heavy metals in the river estuarine sediments–A case study: mahanadi
Feria, J., Marrugo, J., Gonzalez, H., 2010. Heavy metals in Sinú river, department of
basin, India. J. Hazard. Mater. 186, 1837–1846.
Córdoba, Colombia, South America. Rev. Fac. Ing. Univ. Antioq. 55, 35–44.
Taylor, S.R., 1964. Abundance of chemical elements in the continental crust: a new table.
Gupta, N., Khan, D.K., Santra, S.C., 2012. Heavy metal accumulation in vegetables grown
Geochim. Cosmochim. Acta 28, 1273–1285.
in a long-term wastewater-irrigated agricultural land of tropical India. Environ.
Teng, Y.G., Wu, J., Lu, S.J., Wang, Y.Y., Jiao, X.D., Song, L.T., 2014. Soil and soil
Monit. Assess. 184, 6673–6682.
environmental quality monitoring in China: a review. Environ. Int. 69, 177–199.
Hedrick, J.B., 1995. The global rare-earth cycle. J. Alloy. Compd. 225, 609–618.
UPME-MME-UC, 2016. Unidad de Planeación Minero Energética, Ministerio de Minas y
IGAC (Instituto Geográfico Agustín Codazzi). 1979. Métodos analíticos del laboratorio de
Energía, Universidad de Córdoba. Identificación y caracterización de Unidades
suelos. Cuarta edición. Bogotá. pp. 22-55.
Básicas de Beneficio Aurifero en 261 Municipios de Colombia. Informe Técnico
Jiao, W., Chen, W., Chang, A.C., Page, A.L., 2012. Environmental risks of trace elements
(Reporte Final). Bogotá D.C.
associated with long-term phosphate fertilizers applications: a review. Environ.
USEPA, 2007a. SW-846 Test Method 3051A: Microwave Assisted Acid Digestion of
Pollut. 168, 44–53.
Sediments, Sludges, Soils, and Oils. p. 30.
Kabata-Pendias, A., 2011. Trace Elements in Soils and Plants 4th ed.. CRC Press, Boca
USEPA, 2007b. SW-846 Test Method 7471B: Mercury in Solid or Semisolid Waste
Ratón, FL, USA., 505.
(Manual Cold-Vapor Technique). p. 11.
Kelepertzis, E., 2014. Accumulation of heavy metals in agricultural soils of
Viloria de la Hoz, J., 2004. La economía del departamento de Córdoba: ganadería y
Mediterranean: Insights from Argolida basin, Peloponnese, Greece. Geoderma 221–
minería como sectores claves. Documentos de trabajo sobre Economia Regional
222, 82–90.
Banco de la República
Lans, E., Marrugo, J., Ponguta, B., 2008. Study of contamination by organochlorine
Xia, K., Skyllberg, U.L., Bleam, W.F., Bloom, R.P., Nater, E.A., Helmke, P.A., 1999. X-ray
pesticides in the cienaga Grande waters of the low Sinú River valley. Temas Agrar. 13
absorption spectroscopic evidence for the complexation of Hg(II) by reduced sulfur
(1), 49–56.
in soil humic substances. Environ. Sci. Technol. 33, 257–261.
Lasat, M., 2002. Phytoextraction of toxic metals: a review of biological mechanisms. J.
Zhang, H., Feng, X.B., Larssen, T., Shang, L.H., Li, P., 2010. Bioaccumulation of
Environ. Qual. 31, 109–120.
methylmercury versus inorganic mercury in rice (Oryza sativa L.) grain. Environ. Sci.
Li, X., Feng, L., 2012. Multivariate and geostatistical analyzes of metals in urban soil of
Technol. 44, 4499–4504.
Weinan industrial areas, Northwest of China. Atmos. Environ. 47, 58–65.
Zhao, L., Anderson, C.W.N., Qiu, G., Meng, B., Wang, D., Feng, X., 2016a. Mercury
Liu, G., Tao, L., Liu, X., Hou, J., Wang, A., Li, R., 2013. Heavy metal speciation and
methylation in paddy soil: source and distribution of mercury species at a Hg mining
pollution of agricultural soils along Jishui River in non-ferrous metal mine area in
area, Guizhou Province. China Biogeosci. 13, 2429–2440.
Jiangxi Province, China. J. Geochem. Explor. 132, 156–163.
Zhao, L., Qiu, G., Anderson, C.W.N., Meng, B., Wang, D., Shang, L., Yan, H., Feng, X.,
Liu, J., Zhang, X., Tran, H., Wang, D., Zhu, Y., 2011. Heavy metal contamination and risk
2016b. Mercury methylation in rice paddies and its possible controlling factors in the
assessment in water, paddy soil, and rice around an electroplating plant. Environ.
Hg mining area, Guizhou province, Southwest China. Environ. Pollut. 215, 1–9.
Sci. Pollut. Res. 18, 1623–1632.
Zsefer, P., Glasby, G.P., Sefer, K., Pempkowiak, J., Kaliszan, R., 1996. Heavy-metal
Lu, Y., Song, S., Wang, R., Liu, Z., Meng, J., Sweetman, A.J., Jenkins, A., Ferrier, R.C., Li,
pollution in superficial sediments from the southern Baltic Sea off Poland. J.
H., Luo, W., 2015b. Impacts of soil and water pollution on food safety and health
Environ. Sci. Health A 31, 2723–2754.
risks in China. Environ. Int. 77, 5–15.

388