Arch Environ Contam Toxicol (2018) 74:121–130
https://doi.org/10.1007/s00244-017-0459-9
Biomagnification of Mercury in Fish from Two Gold
Mining‑Impacted Tropical Marshes in Northern Colombia
José Luis Marrugo‑Negrete1   · Javier Alonso Ruiz‑Guzmán1 ·
Ana Carolina Ruiz‑Fernández2 
Received: 23 June 2017 / Accepted: 26 September 2017 / Published online: 17 October 2017
© Springer Science+Business Media, LLC 2017
Abstract  Total mercury concentrations (T-Hg) and stable
isotope ratios (δ15N and δ13C) were analyzed in the muscle
of fish, collected at two marshes impacted by gold mining
in Northern Colombia, to evaluate the seasonal and spatial
behavior of these variables and mercury biomagnification
based on their relationship with δ15N ratios in the fish. There
was not significant (p > 0.05) seasonal differences (dry and
rainy season) in δ15N and δ13C values in none marsh, sug-
gesting that these fish species keep a similar feed pattern
within each marsh along the year. However, there were sig-
nificant differences (p < 0.05) between marshes in δ15N,
δ13C, and T-Hg, as well as significant seasonal differences
(p < 0.05) in T-Hg values for some species, suggesting the
possible influence of factors, such as differences in length of
food chains between marshes and different degree of anthro-
pogenic impact in relation to human settlements, farmland,
livestock (higher in Ayapel marsh), and gold mining activity
(higher in La Raya marsh). Correlations analysis showed
significant (p < 0.05) relationships between T-Hg concentra-
tions and δ15N values, as well as trophic level, evidencing a
T-Hg biomagnification process in the fish food chain from
both marshes and consequently a potential health risk for
human riverside inhabitants who eat these fish.
* José Luis Marrugo‑Negrete
jmarrugo@correo.unicordoba.edu.co
1
Water, Applied and Environmental Chemistry Group.
Laboratory of Toxicology and Environmental Management,
University of Córdoba, Cra 6 # 76 – 103, Montería 230002,
Colombia
2
Ana Carolina Ruiz‑Fernández. Instituto de Ciencias del
Mar y Limnología, Universidad Nacional Autónoma de
México, Calz. J. Montes Camarena s/n, Col. Playa Sur,
82040 Mazatlán, Mexico
Wetland areas, including rivers, lakes, marshes, rice fields,
and coastal areas, provide relevant ecosystem services that
contribute to poverty alleviation and human well-being, such
as water availability and fish supply. Fisheries are sometimes
the primary source of animal protein to rural communities
living near wetlands, for which people are highly dependent
on wetland services and are directly harmed by their deg-
radation (Millennium Ecosystem Assessment MEA 2005).
Mercury (Hg) is one of the metals with the highest impact
on aquatic ecosystems due to biotic exposure to methylmer-
cury (MeHg), a highly toxic ionic compound derived from
biomethylation of Hg that bioaccumulates and biomagnifies
in the food chain (Wiener and Suchanek 2008; Marrugo-
Negrete et al. 2010; UNEP 2013). Because Hg biomagni-
fies rapidly, leading to high concentrations in top predators
in aquatic ecosystems, the concentrations of Hg in fish is
strongly influenced by food web structure (Kidd et al. 2003;
Campbell et al. 2008) and food chain length (Tadiso et al.
2011). The stable nitrogen (δ15N) and carbon (δ13C) iso-
topes have been successfully applied worldwide in studies
of trophic structure and pollutant transference in freshwa-
ter (Campbell et al. 2003a, b, 2008; Sampaio da Silva et al.
2005; Tadiso et al. 2011) and marine (Hobson et al. 2002;
Marín-Guirao et al. 2008; Payne and Taylor 2010) ecosys-
tems. Typically, δ15N values have been used to characterize
relative trophic position within the food web, whereas δ13C
values have been used to determine the sources and flow of
carbon transferred from prey to predator (Post 2002; Maraun
et al. 2011).
In Colombia, it has been estimated that between 80 and
100 tons of Hg are released into the atmosphere each year,
from the large gold mining area located at Northern Antio-
quia and Southern Bolivar. Improper mining practices and
waste disposal have resulted in contamination of rivers,
marshes, and other environmental compartments, affecting
13
Vol.:(0123456789)

122
surrounding populations (Marrugo-Negrete et al. 2008a, b;
Olivero-Verbel et al. 2011). The purpose of this study was
to evaluate the seasonal variations (dry and rainy seasons)
of total mercury concentration (T-Hg) in the muscle of fish
from two Colombian marshes (Ayapel and La Raya), which
have been subjected to the impact of gold mining activi-
ties for over two decades. We have used the stable isotopes
of carbon and nitrogen (δ15N, δ13C) to evaluate changes in
feeding habits and the T-Hg biomagnification in the food
chain.
Materials and Methods
Study Area
The Ayapel marsh is located in Northwestern Colombia (8°
17′–8° 24′N and 75° 4′–75° 9′W; Fig. 1) over an approximate
area of 150 km2 at 25 m above sea level. The area has a mean
Fig. 1  Location of the Ayapel and La Raya marshes in NW Colombia
13
Arch Environ Contam Toxicol (2018) 74:121–130
annual precipitation of 2260 mm and mean annual temper-
ature 27 °C. This ecosystem has a great commercial and
ecological importance in Colombia, as an important source
of fish species for human consumption. The Ayapel marsh
receives wastewaters from gold mining activities, which are
discharged in the streams that provide water to the marsh,
and in the flooding season (raining season), it receives waters
from the San Jorge and Cauca rivers, in which intensive
gold mining activities take place. Conversely, the marsh also
receives the runoff from large extensions of farmland and
livestock, as well as domestic wastewater discharges from
the Ayapel municipality and other surrounding human set-
tlements (about 46,000 inhabitants; Departamento Admin-
istrativo Nacional de Estadística de Colombia-DANE 2017).
La Raya marsh is part of the complex from Grande Marsh
of Achí, located in Northwestern Colombia (8° 16′–8° 22′N
and 74° 26′–74° 32′W), with an approximate area of 38 km2
at 30 m above sea level (Fig. 1). The area has a mean annual
precipitation of 2350 mm and a mean temperature of 27 °C.
Arch Environ Contam Toxicol (2018) 74:121–130 123
The marsh is part of Cauca River basin and is located within
the largest gold production area in Colombia. La Raya marsh
is surrounded by the small fishermen village of La Raya
(less than 1000 inhabitants) from which receives minor dis-
charges of domestic, agriculture, and livestock wastewaters
(Marrugo-Negrete et al. 2008a).
Sampling
The fish were captured in Ayapel and La Raya marshes by
using gill nets during January to March (dry season) and
September to November (rainy season) in 2010. Three to
six fish for each species were captured and measured for
standard length (cm). A section of dorsal muscle tissue,
devoid of skin and bones, was removed and kept frozen until
laboratory analysis. The fish species studied were: Hoplias
malabaricus (moncholo), Sorubim cuspicaudus (blanquillo),
Leporinus muyscorum (liseta), and Prochilodus magdale-
nae (bocachico). These species are commonly found in both
marshes and consumed by the local population.
Isotopic Analyses
After freeze drying, the fish samples were ground and then
burned under vacuum while exposed to copper oxide and
pure copper to favor production of the gases for analysis
­(CO2 and ­N2). These gases were recovered by succes-
sive trappings on a vacuum line, according to the method
described by Mariotti (1982), to be analyzed with a PDZ
Europa ANCA-GSL elemental analyzer interfaced to a PDZ
Europa 20–20 isotope ratio mass spectrometer; the analy-
sis was conducted at University of California—Davis. The
stable isotopes contents were expressed as the relative dif-
ference between the samples and international standards.
The delta (δ) unit is given in parts per thousand (‰). The
Pee Dee belemnite carbonate and atmospheric nitrogen were
used as internationals standard values for carbon and nitro-
gen analyses respectively. δX = [(Rsample/Rstandard) − 1] × 103.
Where: X = δ13C or δ15N and R = δ13C/δ12C or δ15N/δ14N.
Trophic Level Calculation
The trophic level (TL) of the evaluated fish species was esti-
mated according to model proposed by Hobson et al. (2002):
TL = TLreference primary consumer
(𝛿 15 Nconsumer − 𝛿 15 Nreference primary consumer )
+
𝛿 15 Nenrichment between TL
where ­TLreference primary consumer is the assumed trophic level
(TL 2) of the reference primary consumer, from which the
relative trophic level of the consumers will be calculated.
δ15Nconsumer and δ15Nreference primary consumer are the mean
nitrogen isotope ratios (‰) of the consumer of interest
and of the reference primary consumer respectively. The
δ15Nenrichment between TL represents the δ15N enrichment pro-
portion between trophic levels.
Mercury Analysis
T-Hg concentrations on wet weight basis (µg/g, ww) were
determined by cold vapor atomic absorption spectropho-
tometry (Thermo Electron AAS series S) after acid diges-
tion with ­H2SO4 and H ­ NO3 (2:1 v/v) for 3 h at 100–110 °C,
according to method described by Sadiq et al. (1991). Accu-
racy of method was determined by analysis of certified refer-
ence material (DORM-2, dogfish muscle from the National
Research Council of Canada, NRCC). T-Hg concentration
measured in DORM-2 was 4.54 ± 0.05 µg/g dry weight (dw)
(certified value 4.64 ± 0.26 µg/g dw), with a mean recovery
of 97.8% (n = 6). Precision of method and RSD were always
< 10%. The limit of detection (3 × standard deviation of 10
blank samples) was 14 ng/g.
A Hazard Index (HI) approach was used to assess the risk
of deleterious effects of methylmercury (MeHg) exposure
through fish consumption (USEPA 1989). HI is defined as
the ratio of the level of exposure to a single substance (E) to
a reference dose (RfD), the latter calculated as the concen-
tration of a particular chemical that is assumed to be safe to
ingest on a daily basis during a lifetime. When HI exceeds
unity, there may be concern for potential health effects, in
particular, for pregnant women and children, who are prone
to develop neurological damage due to MeHg ingestion. The
RfD for MeHg is 0.1 µg/kg/day. Exposure due to MeHg con-
sumption was obtained using the formula, E = (C × I)/W,
where C is the MeHg mean concentration in fish, I is the fish
daily consumption (g/day), and W is the mean weight for a
normal adult person (70 kg). Most freshwater fish contami-
nation studies have shown that MeHg can account for (on
average) more than 90% of the T-Hg concentration in preda-
tor fish tissue (Stahl et al. 2009). Other studies have reported
lower MeHg percentages, ranging from 50 to 72% (Jin et al.
2006; Mason et al. 2006; Riget et al. 2000) and 50–98%
(Carrasco et al. 2011a, b). The amount of MeHg in the fish
species evaluated in this study has been previously reported
between 90.3 ± 2.7 and 93.8 ± 1.9% of the T-Hg concentra-
tion (Marrugo-Negrete et al. 2008b). We have calculated C
as 0.80 × T-Hg for not overestimate exposure.
Statistical Analyses
After checking the normality and homoscedasticity (Shap-
iro–Wilk and Bartlett tests) of the error term, ANOVA and
Tukey’s tests were conducted to compare the differences in
T-Hg concentrations and isotopic ratios (δ15N and δ13C) of
13

124 Arch Environ Contam Toxicol (2018) 74:121–130

fish between seasons and marshes. Pearson’s correlation and
regression analyses were conducted to evaluate the relation-
ship between variables measured. For all statistical tests,
p < 0.05 was considered significant. Computations were
performed using Statistical Analysis Software (SAS) ver-
sion 9.2 and Minitab version 16.0 software.
Results
The standard length, δ15N, δ13C, and T-Hg concentration
values, as well as the estimated relative TL of the fish spe-
cies collected during dry and rainy season in the two marshes
evaluated are shown in Table 1. The average trophic enrich-
ment between carnivorous and noncarnivorous fish for 15N
was of 3.03‰ in La Raya marsh and 3.19‰ in Ayapel marsh,
and for 13C was of 2.04‰ in La Raya marsh and 1.28‰ in
Ayapel marsh.
P. magdalenae, a species from the basis of the trophic chain
that feeds from organic detritus and microalgae (Maldonado-
Ocampo et al. 2005), was considered as the reference primary
consumer, for which was assumed a TL = 2 according to Hob-
son et al. (2002). Considering the δ15N content of P. magdale-
nae (6.57‰ in La Raya marsh and 6.84‰ in Ayapel marsh)
and 15N enrichment between trophic levels observed in this
study (3.03‰ in La Raya marsh and 3.19‰ in Ayapel marsh),
the TL for other fish species was estimated as:
(𝛿 15 Nconsumer − 6.570∕00
TL = 2 + in La Raya marsh, and
3.030∕00
(𝛿 15 Nconsumer − 6.840∕00
=2+ in Ayapel marsh
3.190∕00
There were no significant differences in length, TL, δ15N,
and δ13C values between seasons for almost none of the

Table 1  Stable isotope ratios (δ15N and δ13C), T-Hg concentration and trophic level for fish from Ayapel and La Raya marshes (NW Colombia)
during dry (DS) and rainy (RS) seasons
Marsh Species Season Mean ± SD Trophic level n
15 13
δ N (‰) δ C (‰) T-Hg (µg/g) Length (cm)

La Raya H. malabaricus (C) DS 10.48 ± 0.59 − 27.82 ± 1.36 0.99 ± 0.35* 30.9 ± 6.5 3.3 ± 0.2 6

RS 10.18 ± 0.72 − 28.36 ± 0.92 0.66 ± 0.16* 28.7 ± 4.7 3.2 ± 0.2 5
S. cuspicaudus (C) DS 10.59 ± 0.18 − 27.92 ± 0.79 1.09 ± 0.28* 36.1 ± 4.7 3.3 ± 0.1 5
RS 10.63 ± 0.30 − 27.61 ± 0.85 0.60 ± 0.19* 36.2 ± 6.7 3.3 ± 0.1 4
L. muyscorum (NC) DS 8.49 ± 0.09 − 28.55 ± 0.69 0.34 ± 0.06 26.3 ± 5.1 2.6 ± 0.03 4
RS 8.31 ± 0.14 − 28.92 ± 0.51 0.25 ± 0.05 27.0 ± 4.2 2.6 ± 0.05 4
P. magdalenae (NC) DS 6.73 ± 0.82 − 31.34 ± 0.53 0.19 ± 0.09 25.8 ± 3.8 2.0 ± 0.3 5
RS 6.37 ± 0.62 − 30.78 ± 0.94 0.12 ± 0.02 24.1 ± 4.1 1.9 ± 0.2 4
Ayapel H. malabaricus (C) DS 10.80 ± 0.74 − 29.60 ± 0.61 0.50 ± 0.17 31.1 ± 6.5 3.2 ± 0.2 5
RS 10.67 ± 0.51 − 29.81 ± 0.40 0.33 ± 0.12 27.7 ± 6.3 3.2 ± 0.2 5
S. cuspicaudus (C) DS 11.19 ± 0.45 − 28.76 ± 0.62 0.61 ± 0.14 40.2 ± 4.7 3.4 ± 0.1 5
RS 11.12 ± 0.53 − 28.86 ± 0.58 0.50 ± 0.09 38.2 ± 4.9 3.3 ± 0.2 5
L. muyscorum (NC) DS 8.94 ± 0.28 − 28.62 ± 1.35 0.24 ± 0.07 26.0 ± 5.3 2.7 ± 0.1 4
RS 8.93 ± 0.32 − 29.11 ± 0.74 0.16 ± 0.05 26.7 ± 9.9 2.7 ± 0.1 3
P. magdalenae (NC) DS 7.16 ± 0.93 − 33.19 ± 1.39* 0.13 ± 0.02 23.1 ± 1.7 2.1 ± 0.3 4
RS 6.58 ± 0.98 − 30.80 ± 1.27* 0.10 ± 0.01 20.3 ± 2.4 1.9 ± 0.3 5
Combined data**
 La Raya Mean value (C) DS 10.53 ± 0.44 − 27.87 ± 1.09 1.02 ± 0.31* 33.3 ± 6.1 a 11
RS 10.38 ± 0.60 − 28.03 ± 0.92 0.64 ± 0.16* 32.1 ± 6.6 a 9
Mean value (NC) DS 7.51 ± 1.09 − 30.10 ± 1.57 0.26 ± 0.10 26.0 ± 4.1 b 9
RS 7.34 ± 1.12 − 29.85 ± 1.22 0.19 ± 0.07 25.5 ± 4.2 b 8
 Ayapel Mean value (C) DS 10.99 ± 0.61 − 29.18 ± 0.73 0.56 ± 0.16 35.7 ± 7.2 a 10
RS 10.89 ± 0.55 − 29.33 ± 0.69 0.42 ± 0.14 32.9 ± 7.7 a 10
Mean value (NC) DS 8.05 ± 1.14 − 30.90 ± 2.75 0.19 ± 0.07 24.6 ± 3.9 b 8
RS 7.46 ± 1.44 − 30.17 ± 1.36 0.13 ± 0.04 22.7 ± 6.5 b 8

DS dry season, RS rainy season, n sample size, C carnivorous, NC noncarnivorous fish

*Statistically significant difference between seasons (DS and RS)
**Data in this section correspond to pooled values of C and NC fish

13
Arch Environ Contam Toxicol (2018) 74:121–130 125

species collected at both marshes, except for P. magdalenae, Significant positive correlations were recorded between
in the Ayapel marsh, which showed more negative δ13C val- length of fish with T-Hg, δ15N, and δ13C values in differ-
ues during dry season. Hg concentration in the carnivorous ent species; however, the analysis of pooled data showed
fish species H. malabaricus and S. cuspicaudus from La significant positive correlations between all variables for
Raya marsh were significantly higher during dry season than carnivorous fish and between almost all variables for non-
during rainy season. The comparison of pooled data for car- carnivorous fish in both marshes.
nivorous and noncarnivorous fish between seasons showed Figure 2 shows the relationship between log (T-Hg con-
no significant differences in length, δ15N, and δ13C values, centration) and δ15N values, whereas Fig. 3 shows the rela-
although the mean T-Hg concentrations of carnivorous fish tionship between T-Hg concentration and trophic level for
were significantly higher than in noncarnivorous fish from different fish species from Ayapel and La Raya marshes.
La Raya marsh.
The statistical analysis of pooled data (dry + rainy sea-
sons; Table 2) for each species by sampling site showed that Table 3  Hazard index (HI) for noncarcinogenic effects of T-Hg
δ 15N of S. cuspicaudus from Ayapel marsh were signifi- based on a fish consumption of 120  g/day by people from La Raya
cantly higher than in La Raya marsh. H. malabaricus and and Ayapel marshes
S. cuspicaudus had more negative values of δ13C in Ayapel Marsh Specie HI (mean ± SD)
marsh than in La Raya marsh, whereas T-Hg concentrations
La Raya H. malabaricus 11.6 ± 4.4
were higher in La Raya that in Ayapel in these same species.
S. cuspicaudus 12.0 ± 5.0
When the data were compared according to feeding habits,
L. muyscurum 4.0 ± 0.9
the carnivorous fish had usually higher values of δ 15N, T-Hg
P. magdalenae 2.2 ± 1.0
and length than noncarnivorous fish in both marshes.
Ayapel H. malabaricus 5.7 ± 2.3
Hazard Index for noncarcinogenic effects of T-Hg, based
S. cuspicaudus 7.6 ± 1.7
on a fish consumption of 120  g/day by people from La
L. muyscurum 2.8 ± 0.9
Raya and Ayapel marshes, is presented in Table 3. In both
P. magdalenae 1.6 ± 0.3
marshes, the lowest HI values were determined for the non-
Combined data*
carnivorous fish species P. magdalenae, whereas the highest
 La Raya 7.8 ± 5.5 (1.4–21.0)
values were found for the carnivorous specie S. cuspicaudus.
 Ayapel 4.7 ± 2.9 (1.3–10.6)
Pearson’s correlation coefficients between the variables
measured at each fish species are presented in Table  4. *Minimum and maximum values are presented in parentheses

Table 2  Stable isotope ratios (δ15N and δ13C), T-Hg concentration and trophic level for fish species from Ayapel and La Raya marshes (NW
Colombia)
Species Marsh Mean ± SD n
15 13
δ N (‰) δ C (‰) T-Hg (µg/g) Length (cm) Trophic level

H. malabaricus (C) La Raya 10.34 ± 0.64 b − 28.07 ± 1.16 cd 0.84 ± 0.32 a 29.9 ± 5.6 b 3.3 ± 0.2 a 11

Ayapel 10.73 ± 0.60 ab − 29.71 ± 0.50 b 0.42 ± 0.17 bc 29.4 ± 6.3 bc 3.2 ± 0.2 a 10
S. cuspicaudus (C) La Raya 10.61 ± 0.22 b − 27.78 ± 0.78 d 0.87 ± 0.34 a 36.2 ± 5.3 a 3.3 ± 0.1 a 9
Ayapel 11.16 ± 0.46 a − 28.81 ± 0.57 c 0.56 ± 0.13 b 39.2 ± 4.7 a 3.3 ± 0.1 a 10
L. muyscorum (NC) La Raya 8.40 ± 0.15 c − 28.74 ± 0.60 c 0.29 ± 0.07 cd 26.6 ± 4.3 bc 2.6 ± 0.05 b 8
Ayapel 8.94 ± 0.27 c − 28.83 ± 1.08 bc 0.21 ± 0.07 d 26.3 ± 6.8 bcd 2.7 ± 0.1 b 7
P. magdalenae (NC) La Raya 6.57 ± 0.72 d − 31.09 ± 0.74 a 0.16 ± 0.07 d 25.0 ± 3.8 cd 2.0 ± 0.2 c 9
Ayapel 6.84 ± 0.95 d − 31.86 ± 1.76 a 0.12 ± 0.02 d 21.6 ± 2.5 d 2.0 ± 0.3 c 9
Mean values (C) La Raya 10.46 ± 0.51 a − 27.94 ± 0.99 c 0.86 ± 0.32 a 32.7 ± 6.2 a 20
Ayapel 10.95 ± 0.57 a − 29.25 ± 0.69 b 0.49 ± 0.16 b 34.3 ± 7.4 a 20
Mean values (NC) La Raya 7.43 ± 1.07 b − 29.98 ± 1.38 ab 0.22 ± 0.09 c 25.8 ± 4.0 b 17
Ayapel 7.76 ± 1.29 b − 30.54 ± 2.13 a 0.16 ± 0.06 c 23.6 ± 5.3 b 16
General La Raya 9.07 ± 1.73 b − 28.88 ± 1.56 b 0.57 ± 0.40 a 29.5 ± 6.3 a 37
Ayapel 9.53 ± 1.86 a − 29.83 ± 1.62 a 0.34 ± 0.21 b 29.6 ± 8.4 a 36

C carnivorous, NC noncarnivorous fish

Letters “a” and “d” correspond to the highest and lowest value, respectively; mean values that share at least one letter are not significantly differ-
ent

13

126 Arch Environ Contam Toxicol (2018) 74:121–130

Table 4  Pearson’s correlation coefficients for the relationships level to another (Hobson and Welch 1992; Michener and
between length, stable isotope ratios (δ15N and δ13C), and T-Hg con- Schell 1994; Middelburg 2014). The results obtained from
centrations in fish from Ayapel and La Raya marshes
this study agree with this pattern, where mean isotopic
Marsh Species Variable Length (cm) δ15N δ13C enrichment of 3.03 and 3.19‰ in δ15N, and 2.04 and 1.28‰
La Raya H. malabaricus (C) δ15 N 0.76**
in δ13C, were recorded between noncarnivorous and car-
δ13 C 0.57 0.72* nivorous fish in La Raya and Ayapel marshes respectively.
T-Hg 0.89** 0.65* 0.54 Other studies have found the same increase pattern with
S. cuspicaudus (C) δ15 N 0.57 isotopic enrichments around 2–3‰ in δ15N and 1–2‰ in
δ13 C 0.86** 0.40 δ13C between noncarnivorous and carnivorous fish (Sam-
T-Hg 0.56 0.30 0.30 paio da Silva et al. 2005; Gustin et al. 2005). The TL results
L. muyscurum (NC) δ15 N 0.40 showed a logic assignment of trophic position for the fish
δ13 C 0.78* 0.78* species evaluated, according to their feeding habits. So, P.
T-Hg 0.56 0.78* 0.76* magdalenae, an iliophago/detritivorous species (Restrepo
P. magdalenae (NC) δ15 N 0.86** 2001; Maldonado-Ocampo et al. 2005) is located in a trophic
δ13 C 0.23 0.34
level lower that H. malabaricus and S. cuspicaudus (carnivo-
T-Hg 0.71* 0.83** − 0.22
rous species, Atencio-Garcia et al. 2005; Valderrama et al.
Mean values (C) δ15 N 0.69**
2006), whereas L. muyscurum (omnivorous species, Olaya
δ13 C 0.64** 0.66*
et al. 2011) is located in an intermediate TL.
T-Hg 0.66** 0.50* 0.44*
Mean values (NC) δ15 N 0.48
The lack of significant interseasonal differences in δ15N
δ13 C 0.40 0.85**
and δ13C values of the fish in both marshes suggested that
T-Hg 0.59* 0.86** 0.71* these species keep a similar feeding pattern along the year,
Ayapel H. malabaricus (C) δ15 N 0.89** which also explains their constant TL in both seasons at each
δ13 C 0.47 0.66* marsh (Table 1). However, the differences between marshes
T-Hg 0.93** 0.82** 0.61 in pooled data of δ15N value for S. cuspicaudus and δ13C val-
S. cuspicaudus (C) δ15 N 0.53 ues for H. malabaricus and S. cuspicaudus (Table 2), as well
δ13 C 0.23 0.82** as the significant differences between marshes in the total
T-Hg 0.67* 0.57 0.31 data for these variables (δ15N and δ13C), which, consider-
L. muyscurum (NC) δ15 N 0.82* ing there are not differences in fish length between marshes
δ13 C 0.75 0.89** (Table 2), suggested possible differences in the source of C
T-Hg 0.64 0.56 0.82* and N at each marsh.
P. magdalenae (NC) δ15 N 0.51
The higher δ 15N and TL values in S. cuspucuadus from
δ13 C − 0.56 0.00
Ayapel marsh, as well as the higher δ 15N value for total data
T-Hg 0.72* 0.54 − 0.47
of Ayapel marsh, might be explained by: 1) differences in the
Mean values (C) δ15 N 0.77**
δ13 C 0.64** 0.76**
length of food web between marshes, or 2) an enrichment in
15
T-Hg 0.86** 0.77** 0.60**
N contents derived from a higher input of anthropogenic
Mean values (NC) δ15 N 0.57* sources, such as the discharges of urban waste water and run-
δ13 C 0.41 0.65** off from farm lands and livestock facilities. Previous studies
T-Hg 0.73** 0.72** 0.64** have recognized that the anthropogenic inputs of nitrogen
increased significantly the δ15N values in the aquatic ecosys-
*Significant statistical differences at p < 0.05 tems (McClelland and Valiela 1998; Fry et al. 2003; Vizzini
**Significant statistical differences at p < 0.01 et al. 2005; Piola et al. 2006; Marín-Guirao et al. 2008) even
at small spatial scales (Vizzini and Mazzola 2006). The
In the two marshes and seasons, these relationships had
Ayapel marsh is surrounded by several human settlements
significant correlation coefficients and the regression lines
and large extensions of land dedicated to agriculture and
had positives slopes.
livestock, which represent a potential source anthropogenic
input of nitrogen to the marsh and might explain the higher
δ15N value in S. cuspicuaudus and the total data from Ayapel
Discussion
marsh compared with La Raya marsh. Of note, fluctuations
of the baseline δ15N may hinder the interpretation of trophic
Trophic Level and Stable Isotopes
differences within species when δ15N values are compared
between sampling sites. For this reason, it is important to be
Increments of the order of 2–4‰ in δ15N and 1‰ in δ13C
careful when interpreting changes in the TL based on δ15N
are generally recognized as a transition from one trophic
values (Hobson et al. 2002).

13
Arch Environ Contam Toxicol (2018) 74:121–130 127

Fig. 2  Relationship between 0.2

0.0
nitrogen isotopic ratio (δ15N) Log T-Hg=-2.124+ 0.200 δ15N Log T-Hg=-2.023+ 0.1584
and logarithm of mercury r2 92.1%
0.0 -0.2 δ15N
concentrations (T-Hg) in fish p <0.001 r2 91.8%
from Ayapel and La Raya
-0.2 p <0.001
marshes (Colombia) during dry -0.4
(DS) and rainy (RS) seasons.
H. malabaricus (circle), S. -0.4
-0.6
cuspicuaudus (triangle), L. muy-
scorum (square), P. magdalenae -0.6
(diamond) -0.8
-0.8

-1.0
-1.0
-Hg)

6 7 8 9 10 11 6 7 8 9 10 11 12

0.0 -0.2
Log T-Hg=-2.010+ 0.1722 δ15N Log T-Hg=-1.888+ 0.1343 δ15N
r2 93.3% r2 83.4%
-0.2
p <0.001 -0.4
p <0.001

-0.4 -0.6

-0.6
-0.8

-0.8
-1.0

-1.0
-1.2

5 6 7 8 9 10 11 12 5 6 7 8 9 10 11 12
δ15N

Furthermore, in any aquatic ecosystem, the organic car-
bon become from the two main sources: 1) autochthonous
material, such as phytoplankton and aquatic macrophytes,
and 2) allochthonous terrestrial material, transported by
runoff from the watersheds that might include waste waters
from industrial, domestic, and agricultural activities (Finlay
and Kendall 2007; Harmelin-Viven et al. 2010). It has been
described that the freshwater phytoplankton presents more
negative values of δ13C (− 35 to − 5.9‰, Vuorio et al. 2006)
than detritus of terrestrial matter, composed mainly of veg-
etal remain (− 9 to − 14‰ for C4 plants, and from − 30 to
− 23‰ for C3 plants) (Maksymowska et al. 2000, Finlay
and Kendall 2007; Harmelin-Viven et al. 2010). According
to this, the significant differences between marshes in δ13C
values obtained for H. malabaricus and S. cuspicaudus, as
well as the total data for this variable (Table 2), suggests pos-
sible differences in the carbon proportion provided by each
source (autochthonous and allochthonous) to these ecosys-
tems. In this case, the significantly more negative δ13C val-
ues obtained in fish from Ayapel marsh could be explained
considering the fact that the waste water of agricultural and
livestock activities taking place in the surroundings, pro-
vided great amount of nutrients (mainly nitrogen), stimulat-
ing a higher growth of phytoplankton in the Ayapel marsh
than in La Raya marsh. Another study reported significantly
more negative δ13C values in particulate organic matter
river waters during the spring, which were related to higher
growth of phytoplankton (Harmelin-Viven et al. 2010).
Hg Pollution
The seasonal differences of T-Hg concentrations observed
in H. malabaricus and S. cuspicaudus collected in La Raya
marsh are not attributable to seasonal differences in length
or trophic level of fish, because these variables did not show
significant differences between seasons (Table 1). These
results might be explained by factors as: (1) increased Hg
absorption through gills and other body permeable surfaces,
due to higher T-Hg concentrations in water, as a result of
seasonal changes in water level. In fact, during the dry sea-
son, the water levels in La Raya marsh can be 3.3 m lower
than during the rainy season (Marrugo-Negerete et  al.
2008a); or (2) increased Hg concentration in the food of the
fish for the same reason described above. Previous studies
have shown that even if diet is a major source of Hg for fish,
other sources, such as Hg absorbed trough of gills, ingested
with the water, and the Hg concentrations of the environ-
ment where the organisms live, should be taken into account
when differences in T-Hg concentrations between fish from

13

128 Arch Environ Contam Toxicol (2018) 74:121–130

Fig. 3  Relationship between 1.6 0.8

trophic level (based on δ15N) T-Hg = - 1.316 + 0.698 TL T-Hg = - 0.698 + 0.376 TL
1.4 0.7
and T-Hg concentrations in r2 75.7% r2 81.5%
fish from Ayapel and La Raya 1.2 p <0.001 0.6 p <0.001
marshes during dry (DS) and
rainy (RS) seasons. H. mala- 1.0 0.5
baricus (circle), S. cuspicuau-
0.8 0.4
dus (triangle), L. muyscorum
(square), P. magdalenae 0.6 0.3
(diamond)
0.4 0.2

0.2 0.1

0.0 0.0

2.0 2.5 3.0 3.5 2.0 2.4 2.8 3.2 3.6

0.9
0.6
T-

0.8 T-Hg = - 0.710 + 0.406 TL T-Hg = - 0.389 + 0.243 TL

r2 83.5% 0.5 r2 72.1%
0.7
p <0.001 p <0.001
0.6 0.4
0.5
0.3
0.4

0.3 0.2

0.2
0.1
0.1

0.0 0.0

1.5 2.0 2.5 3.0 3.5 1.5 2.0 2.5 3.0 3.5

Trophic level

different zones are evaluated (Campbell et al. 2003a, b; Dur-
rieu et al. 2005; Sharma et al. 2009).
In addition, the higher Hg concentration of H. malabari-
cus and S. cuspicaudus in La Raya marsh might be explained
by a largest anthropogenic contribution of Hg derived from
the gold mining activity that takes place in the surroundings
of this marsh. Previous studies in this zone have evidenced
a higher impact from the gold mining activities in La Raya
marsh than in Ayapel marsh (Marrugo et al. 2007; Marrugo-
Negrete et al. 2008a, b, 2010; Olivero-Verbel et al. 2011).
However, the fish from both marshes are polluted with
mercury, and according to HIs estimated based on a daily
intake of fish of 120 g (Table 3), there is a potential health
risk for riverside inhabitants who consume them, mainly
the carnivorous fish. These results agreed with previous
studies in the Ayapel marsh (HI = 0.5–10, Marrugo et al.
2007; HI = 5.7 ± 2.1, Marrugo-Negrete et al. 2008b) and
the marshes complex (Grande marsh) to which it belongs La
Raya marsh (HI = 1.2–31.5, Marrugo-Negrete et al. 2008a),
suggesting that a health risk has existed for several years.
Hg Biomagnification
The significant correlations observed between T-Hg con-
centrations and fish length for some fish species, as well as
pooled data for carnivorous and noncarnivorous fish in both
marshes (Table 4), suggested a highest Hg bioaccumula-
tion in the older fish due to longer exposure time. Increases
in T-Hg concentrations with length have been previously
reported for these same species in Ayapel marsh and other
marshes of the Mojana region-Colombia (Marrugo et al.
2007; Marrugo-Negrete et al. 2008a, b, 2010).
The significant correlations observed between δ15N
values and T-Hg concentrations for some fish species and
for pooled data in carnivorous and noncarnivorous fish
(Table 4), as well as the significant correlation observed
between the log T-Hg concentrations with δ15N values and
the TL (Figs. 2, 3), suggested a biomagnification process
within the fish food chain at both marshes. Other studies also
have reported significant correlations between T-Hg con-
centrations and δ15N values in freshwater fish as Hemiodus
unimaculatus, Pygocentrus nattereri (Sampaio da Silva et al.
2005), catfish Clarias gariepinus, and golden carp Carassius
auratus (Tadiso et al. 2011), among others.
The regression slope obtained from the regression analy-
sis between log T-Hg concentrations and δ15N values can be
used as an index of the biomagnification rate (Campbell et al.
2003a, b, 2005a, b, 2008; Tadiso et al. 2011). In this study,
the slopes ranged from 0.13 to 0.20 between seasons and
marshes (Fig. 2), indicating a T-Hg biomagnification process

13
Arch Environ Contam Toxicol (2018) 74:121–130 129
within the fish food chain at both marshes. Similar results
have been recorded for food chains of freshwater fishes from
other places, such as the Tanganyka Lake (slope = 0.22,
Campbell et al. 2008), Ziway Lake (slope = 0.13, Tadiso
et al. 2011), Victoria and Malawi Lakes (slope = 0.12–0.20),
and globally (Campbell et al. 2003a, b, 2005a, b; Kidd et al.
2003).
Conclusions
The relationship observed between T-Hg concentrations and
δ15N values, as well as the analysis of the trophic levels,
evidenced a T-Hg biomagnification process in the fish food
chain from the Ayapel and La Raya marshes. The differ-
ences observed in the values of δ15N and δ13C in the fish
collected from both marshes may be mainly related to the
anthropization status of each site, because the Ayapel marsh
has a higher development of human settlements and agri-
cultural activities than La Raya marsh. The higher T-Hg
concentration found in fish from La Raya marsh is indica-
tive of a higher impact from the gold mining activity in this
area. The evidence of mercury-contaminated fish and the
potential risk to the health of residents in the area deserve
careful risk–benefit considerations for fish consumption and
attention should be addressed to minimize the exposure to
mercury in both ecosystems.
Acknowledgments  The authors thank the Administrative Department
of Science, Technology, and Innovation of Colombia COLCIENCIAS
for financial support (Grant 223-2010).
References
Atencio-Garcia V, Kerguelén-Durango E, Cura E, Rosado R, Vallejo A,
Valderrama M (2005) Régimen alimentario de siete especies ícti-
cas en el embalse de la hidroeléctrica de Urrá (Córdoba, Colom-
bia). Rev MVZ Córdoba 10(2):614–622
Campbell LM, Dixon DG, Hecky RE (2003a) A review of mercury in
Lake Victoria, East Africa: implications for human and ecosystem
health. J Toxicol Environ Health B Crit Rev 6:325–356
Campbell LM, Hecky RE, Nyaundi J, Muggide R, Dixon DG (2003b)
Distribution and food-web transfer of mercury in Napoleon and
Winam Gulfs, Lake Victoria East Africa. J Gt Lakes Res 29(Suppl
2):267–282
Campbell LM, Hecky RE, Dixon DG, Thacker RJ, Wandera SB
(2005a) Trophic niche segregation in the Nilotic ichthyofauna of
Lake Albert (Uganda). Environ Biol Fishes 74:247–260
Campbell LM, Norstrom RJ, Hobson KA, Muir DCG, Backus S, Fisk
AT (2005b) Mercury and other trace elements in a pelagic arc-
tic marine food web (Northwater Polynya, Baffin Bay). Sci Total
Environ 351–352:247–263
Campbell L, Verburg P, Dixon DJ, Hecky RE (2008) Mercury biomag-
nification in the food web of Lake Tanganyika (Tanzania, East
Africa). Sci Total Environ 402:184–191
Carrasco L, Benejam L, Benito J, Bayona JM, Díez S (2011a) Meth-
ylmercury levels and bioaccumulation in the aquatic food
web of a highly mercury-contaminated reservoir. Environ Int
37:1213–1218
Carrasco L, Barata C, García-Berthou E, Tobias A, Bayona JM, Díez
S (2011b) Patterns of mercury and methylmercury bioaccumula-
tion in fish species downstream of a long-term mercury-contam-
inated site in the lower Ebro River (NE Spain). Chemosphere
84:1642–1649
Departamento Administrativo Nacional de Estadística de Colombia-
DANE (2017) Proyecciones de población municipales por área
2005–2020. http://www.dane.gov.co/index.php/estadisticas-
por-tema/demografia-y-poblacion/proyecciones-de-poblacion.
Accessed 23 June 2017
Durrieu G, Maury-Brachet R, Boudou A (2005) Gold mining and
mercury contamination of the piscivorous fish Hoplias aimara
in French Guiana (Amazon basin). Ecotoxicol Environ Saf
60:315–323
Finlay JC, Kendall C (2007) Stable isotope tracing of temporal and
spatial variability in organic matter sources to freshwater eco-
systems. In: Michener R, Lajtha K (eds) Stable isotopes in ecol-
ogy and environmental science, 2nd edn. Blackwell Publications,
Malden, pp 283–333
Fry B, Gace A, McClelland JW (2003) Chemical indicators of anthro-
pogenic nitrogen loading in four Pacific estuaries. Pacific Sci
57:77–101
Gustin MS, Saitoa L, Peacock M (2005) Anthropogenic impacts on
mercury concentrations and nitrogen and carbon isotope ratios in
fish muscle tissue of the Truckee River watershed, Nevada, USA.
Sci Total Environ 347:282–294
Harmelin-Vivien M, Dierking J, Bănaru D, Fontaine MF, Arlhac D
(2010) Seasonal variation in stable C and N isotope ratios of the
Rhone River inputs to the Mediterranean Sea (2004–2005). Bio-
geochemistry 100:139–150
Hobson KA, Welch HE (1992) Determination of trophic relationships
within a high Arctic marine food web using stable-isotope analy-
sis. Mar Ecol Prog Ser 84:9–18
Hobson KA, Fisk A, Karnovsky N, Holst M, Gagnon J-N, Fortier
M (2002) A stable isotope (δ 13C, δ15N) model for the North
Water food web: implications for evaluating trophodynamics
and the flow of energy and contaminants. Deep Sea Res Part
2(49):5131–5150
Jin LN, Liang LN, Jiang GB, Xu Y (2006) Methylmercury, total mer-
cury and total selenium in four common freshwater fish species
from Ya-Er Lake, China. Environ Geochem 28:401–407
Kidd KA, Bootsma HA, Hesslein RH, Lockhart L, Hecky RE (2003)
Mercury concentrations in the food web of Lake Malawi East
Africa. J Great Lakes Res 29(Suppl. 2):258–266
Maksymowska D, Richard P, Piekarek-Jankowska H, Riera P (2000)
Chemical and isotopic composition of the organic matter sources
in the Gulf of Gdansk (Southern Baltic Sea). Estuar Coast Shelf
Sci 51:585–598
Maldonado-Ocampo JA, Ortega-Lara A, Usma JS, Galvis G, Villa-
Navarro FA, Vásquez L, Prada-Pedreros S, Ardila C (2005) Peces
de los Andes de Colombia. Instituto de investigación de recursos
biológicos Alexander Von Humboldt, Bogotá
Maraun M, Erdmann G, Fischer BM, Pollierer MM, Norton RA,
Schneider K, Schue S (2011) Stable isotopes revisited: their use
and limits for oribatid mite trophic ecology. Soil Biol Biochem
43:877–882
Marín-Guirao L, Lloret J, Marin A (2008) Carbon and nitrogen stable
isotopes and metal concentration in food webs from a mining-
impacted coastal lagoon. Sci Total Environ 393:118–130
Mariotti A (1982) Apports de la geochimie organique a` la connais-
sance du cycle de l’azote. Ph.D. thesis. University Pierre et Marie
Curie, France
13

130
Marrugo J, Lans E, Benítez L (2007) Hallazgo de mercurio en peces
de la Ciénaga de Ayapel, Córdoba Colombia. Rev MVZ Córdoba
12(1):878–886
Marrugo-Negrete J, Benitez L, Olivero J (2008a) Distribution of mer-
cury in several environmental compartments in an aquatic eco-
system impacted by gold mining in Northern Colombia. Arch
Environ Contam Toxicol 55:305–316
Marrugo-Negrete J, Olivero J, Lans E, Benitez L (2008b) Total mer-
cury and methylmercury concentrations in fish from the Mojana
region of Colombia. Environ Geochem Health 30:21–30
Marrugo-Negrete J, Benítez LN, Olivero-Verbel J, Lans E, Vazquez F
(2010) Spatial and seasonal mercury distribution in the Ayapel
Marsh, Mojana region Colombia. Int J Environ Health Res
20(6):451–459
Mason RP, Heyes D, Sveinsdottir A (2006) Methylmercury concen-
trations in fish from tidal waters of the Chesapeake Bay. Arch
Environ Contam Toxicol 51:425–437
McClelland JW, Valiela I (1998) Linking nitrogen in estuarine produc-
ers to land-derived sources. Limnol Oceanogr 43:577–585
Michener RH, Schell DM (1994) Stable isotope ratios as tracers in
marine and aquatic food webs. In: Lajtha K, Michener RH (eds)
Stable isotopes in ecology and environmental science. Blackwell
Scientific Publications, Oxford, pp 138–157
Middelburg JJ (2014) Stable isotopes dissect aquatic food webs from
the top to the bottom. Biogeosciences 11:2357–2371
Millennium Ecosystem Assessment-MEA (2005) Ecosystems and
human well-being: wetlands and water synthesis. A report of the
Millennium Ecosystem Assessment. World Resources Institute,
Washington
Olaya C, Segura F, Tordecilla G, Appeldoorn R (2011) Estimación de
los parámetros biológicos básicos de peces comerciales del río
Sinú-Fase III. Universidad de Córdoba, Monteria
Olivero-Verbel J, Caballero-Gallardo K, Negrete-Marrugo J (2011)
Relationship between localization of gold mining areas and hair
mercury levels in people from Bolivar, North of Colombia. Biol
Trace Elem Res 144:118–132
Payne EJ, Taylor DL (2010) Effects of diet composition and trophic
structure on mercury bioaccumulation in temperate flatfishes.
Arch Environ Contam Toxicol 58:431–443
Piola RF, Moore SK, Suthers IM (2006) Carbon and nitrogen isotope
analysis of three types of oyster tissue in an impacted estuary.
Estuar Coast Shelf Sci 66:255–266
Post DM (2002) Using stable isotopes to estimate trophic position:
models, methods, and assumptions. Ecology 83(3):703–718
Restrepo D (2001) Determinación de algunos aspectos biológicos del
bocachico Prochilodus magdalenae (Steindachner 1878, Char-
acidae, Characiformes), en la zona media y baja de la cuenca del
13
Arch Environ Contam Toxicol (2018) 74:121–130
río Sinú, Córdoba, Colombia. Corporación Autónoma Regional
de los Valles del Sinú y San Jorge-CVS, Montería
Riget F, Asmund G, Aastrup P (2000) Mercury in Arctic char (Salve-
linus alpinus) populations from Greenland. Sci Total Environ
245:161–172
Sadiq M, Zaidi T, At-Mohana H (1991) Sample weight and digestion
temperature as critical factors in mercury determination in fish.
Bull Environ Contam Toxicol 47:335–341
Sampaio da Silva D, Lucotte M, Roulet M, Poirier H, Mergler D,
Oliveira Santos E, Crossa M (2005) Trophic structure and bioac-
cumulation of mercury in fish of three natural lakes of the Brazil-
ian Amazon. Water Air Soil Pollut 165:77–94
Sharma CM, Rosseland BO, Almvik M, Eklo OM (2009) Bioaccumu-
lation of organochlorine pollutants in the fish community in Lake
Arungen, Norway. Environ Pollut 157:2452–2458
Stahl LL, Snyder BD, Olsen AR, Pitt JL (2009) Contaminants in fish
tissue from U.S. lakes and reservoirs: a national probabilistic
study. Environ Monit Assess 150:3–19
Tadiso TM, Borgstrøm R, Rosseland BO (2011) Mercury concentra-
tions are low in commercial fish species of Lake Ziway, Ethiopia,
but stable isotope data indicated biomagnification. Ecotoxicol
Environ Saf 74(4):953–959
United Nation Environmental Protection Agency-USEPA (1989) Risk
assessment guidance for superfund vol 1. Human Health Evalu-
ation Manual (Part A). EPA/540/1-89/002. USEPA, Washington
United Nations Environment Programme-UNEP (2013) Global mer-
cury assessment 2013. Sources, emissions, releases and envi-
ronmental transport. UNEP Chemicals Branch, Geneva. URL
http://www.unep.org/PDF/PressReleases/GlobalMercuryAssess-
ment2013.pdf. Accessed 5 May 2014
Valderrama M, Salas F, Solano D (2006) Los peces y las pesquería en
el embalse de Urrá 2001–2005. Fundación bosques y humedales/
Empresa Urrá S.A. E.S.P/Incoder, Montería
Vizzini S, Mazzola A (2006) The effects of anthropogenic organic
matter inputs on stable carbon and nitrogen isotopes in organisms
from different trophic levels in a southern Mediterranean coastal
area. Sci Total Environ 368:723–731
Vizzini S, Benedetto S, Thang DC, Mazzola A (2005) Spatial vari-
ability of stable carbon and nitrogen stable isotope ratios in a
Mediterranean coastal lagoon. Hidrobiologia 550:73–82
Vuorio K, Meili M, Sarvala J (2006) Taxon-specific variation in the
stable isotopic signatures (δ13 C and δ15 N) of lake phytoplank-
ton. Freshw Biol 51:807–822
Wiener JG, Suchanek TH (2008) The basis for ecotoxicological con-
cern in aquatic ecosystems contaminated by historical mercury
mining. Ecol Appl 18(Suppl. 8):A3–A11