Heavy metal pollution downstream the abandoned Coval da Mó mine (Portugal) and associated effects on epilithic diatom communities.

Abstract

This study examined trace-element concentrations in 39 sediment samples collected in the vicinity of the abandoned Coval da Mó mine, and evaluated the anthropogenic contaminant effects and other environmental variables in the taxonomic composition, structure and morphological changes of benthic diatom communities. The results show the existence of extremely high contamination in Pb, Zn and Cd (the mean values exceed the background values 376, 96 and 19 times, respectively) on the first 2.5 km in the water flow direction. Also Co, Cu, Mn and Ni are present in high concentrations. Dilution by relatively uncontaminated sediment reduces metal concentrations downstream, but Zn concentrations increase downstream Fílvida stream, as a result of several factors such as sewage and agriculture. To evaluate the biological effects caused by Pb, Cd and Zn, three sites were selected. In the stressed environment, near the mining area (C232), diatoms were extremely rare, however there was a slight recovery at site C79 located 2km downstream. Fragilaria capucina var. rumpens, Fragilaria cf. crotonensis and Achnanthidium minutissimum showed abnormal valves which may be related to high levels of metals. Six km downstream, in Fílvida stream (C85), an increase in species richness and diversity was registered while the relative percentage of valve teratologies was lower. In the absence of OM, nutrients and low pH the diatom community patterns must be attributed to the metal concentration at some sites. Considering that community diversity can be affected by abiotic and biotic variables and valve deformations are caused by a small number of variables, basically metals, and acid conditions, we consider the presence of teratologies as an indication of the presence of metals.

Full text

Page 1

Heavy metal pollution downstream the abandoned Coval da Mó mine (Portugal) and
associated effects on epilithic diatom communities
Eduardo Ferreira da Silvaa,⁎, Salomé F.P. Almeidab, Marcelo L. Nunesa, Ana T. Luísa, Fredrik Borgc,
Markus Hedlundc, Carlos Marques de Sád, Carla Patinhaa, Paula Teixeiraa
aGeoBioTec — GeoBioSciences, GeoTechnologies and GeoEngineering Research Center and Department of Geosciences, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal
bGeoBioTec — GeoBioSciences, GeoTechnologies and GeoEngineering Research Center and Department of Biology, University of Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal
cDepartment of Environmental Engineering, Division of Applied Geology, Lulea, University of Technology, Sweden
dDepartment of Geology, Faculty of Sciences, University of Porto. Rua do Campo Alegre 687, 4169-007 Porto, Portugal
a b s t r a c ta r t i c l ei n f o
Article history:
Received 2 March 2009
Received in revised form 1 June 2009
Accepted 29 June 2009
Available online 31 July 2009
Keywords:
Mining
Metals
Biological indicators
Diatoms
Teratologies
This study examined trace-element concentrations in 39 sediment samples collected in the vicinity of the
abandoned Coval da Mó mine, and evaluated the anthropogenic contaminant effects and other environ-
mental variables in the taxonomic composition, structure and morphological changes of benthic diatom
communities.
The results show the existence of extremely high contamination in Pb, Zn and Cd (the mean values exceed
the background values 376, 96 and 19 times, respectively) on the first 2.5 km in the water flow direction. Also
Co, Cu, Mn and Ni are present in high concentrations. Dilution by relatively uncontaminated sediment reduces
metal concentrations downstream, but Zn concentrations increase downstream Fílvida stream, as a result of
several factors such as sewage and agriculture.
To evaluate the biological effects caused by Pb, Cd and Zn, three sites were selected. In the stressed envi-
ronment,neartheminingarea(C232),diatomswereextremelyrare,howevertherewasaslightrecoveryatsite
C79 located 2 km downstream. Fragilaria capucina var. rumpens, Fragilaria cf. crotonensis and Achnanthidium
minutissimum showed abnormal valves which may be related to high levels of metals.
Six km downstream, in Fílvida stream (C85), an increase in species richness and diversity was registered while
the relative percentage of valve teratologies was lower. In the absence of OM, nutrients and low pH the diatom
community patterns must be attributed to the metal concentration at some sites. Considering that community
diversitycan be affected by abiotic and biotic variables and valve deformations are caused by a small number of
variables, basically metals, and acid conditions, we consider the presence of teratologies as an indication of the
presence of metals.
© 2009 Elsevier B.V. All rights reserved.
1. Introduction
Human activities such as mining may greatly increase trace-
element concentrations in the environment (Brigham, 2002). The
wastes from mining activity, containing high metal concentrations,
represent a source of metal contamination for a long time following
extraction. Because of chemical and geotechnical instabilities of these
materials, and other potential environmental constraints, the wastes
result in long-term public concerns. The mining wastes contribute
with sediment, acidity, metals, and secondary precipitates to streams,
which may render local watersheds inhospitable to aquatic biota
(Kucken et al., 1994; Lottermoser et al., 1999; Gold et al., 2002;
Olaveson and Nalewajko, 2000; Sabater, 2000; Hirst et al., 2002;
Hammarstrom et al., 2003; Lacoul and Freedman, 2006).
The traditional approach to investigate the environmental impacts of
anabandonedminingsiteisbasedongeochemicalsurveymethodologies.
Streambed sediments can be a useful medium for trace-element
analyses. Streambed sediments can accumulate chemicals over time,
and may be useful archives of past contamination. If significant trace-
elementcontaminationisintroducedtoastream–eithertransientlyor
continuously– streambed sedimentsshould accumulate someportion
of the elements through chemical and physical sorption processes.
Measurement of metals total concentration in sediments is useful
to detect changes in the stream due to different possible phenomena
such as erosion and leaching to groundwater, but it does not give any
indication about the chemical form of metals in sediments (Pagnanelli
et al., 2004). The knowledge about metal partitioning among different
geochemical phases is particularly important to assess the potentially
bioavailable fractions and any risks of ecotoxicity.
Numerous geochemical studies on total trace metal concentrations
or related to partitioning of trace metals in different geochemical
Science of the Total Environment 407 (2009) 5620–5636
⁎ Corresponding author.
E-mail address: eafsilva@ua.pt (E. Ferreira da Silva).
0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.scitotenv.2009.06.047
Contents lists available at ScienceDirect
Science of the Total Environment
journal homepage: www.elsevier.com/locate/scitotenv

Page 2

phases have been performed in order to study water contamination
by heavy metals after cessation of sulphide-ore mining activities
(Nordstrom, 1982; Leenaers, 1989; Boulet and Larocque, 1998;
Edelgaard and Dahlströ, 1999; Leblanc et al., 2000; Schroth, 2001;
Bird et al., 2003; Ferreira da Silva et al., 2005, 2006; Sánchez España
et al., 2005).
Diatom communities possess many of the attributes required for
indicator organisms because they are widely distributed, occupying
an essential position at the base of aquatic food chains as they are
important primary producers in many freshwater environments.
They may attach to substrates, therefore integrating the real habitat
conditions and responding quicker to environmental changes than
higher level organisms, because of their short life cycle. In addition,
algal assemblages are rich in species, composed of some tens of
taxa with environmental tolerances and preferences (Lange-
Bertalot, 1979; Ter Braak and Van Dam, 1989; Cox, 1991; Van
Dam et al., 1994; Kelly and Whitton, 1989; Pan et al., 1996;
Stoermer and Smol, 1999). The susceptibility of freshwater diatom
communities to metals has been reported under field and
laboratory conditions (Say, 1978; Leland and Carter, 1984; Denise-
ger et al., 1986; Genter et al., 1987; Gray and Hill, 1995; Gustavson
and Wängberg, 1995; Genter, 1996; Medley and Clements, 1998;
Paulsson et al., 2000).
Many studies on metal polluted rivers have shown that diatoms
respond to environmental degradation not only at the community
level through shifts in dominant taxa and diversity patterns but
also at the individual level with changes in frustule morphology.
Size decrease (Gensemer 1990; Cattaneo et al., 1998, 2004) and
frustule deformations (Harding and Whitton, 1976; Thomas et al.,
1980; Adshead-Simonsen et al., 1981; Barber and Carter, 1981;
Foster, 1982; Kelly and Whitton, 1989; Carter, 1990; Yang and
Duthie, 1993; McFarland et al. 1997; Dickman, 1998; Gold et al.,
2003; Cattaneo et al., 2004) have been correlated with high metal
concentrations. An interesting review paper on teratologies in
diatoms has recently been published (Falasco et al. 2009). This
review highlights the main causes for frustule deformations.
Although metal contamination is pointed out as one of the
producers of teratological alterations in diatom frustules it is not
the only one. Other physical and chemical parameters were pointed
as potential deformation causers such as: drought conditions, light
intensity, UV, salinity levels, nutrients and other toxic compounds
such as cyanide, polycyclic aromatic hydrocarbons (PAH) and
pesticides.
Several studies that aim to establish a relationship between the
level of metal contamination and diatom species composition have
been carried out (Besch et al.,1972; Say,1978; Say and Whitton,1980;
Rushforthetal.,1981;Denisegeretal.,1986;GenterandLehman,2000;
Sabater, 2000; Gold et al., 2002), thereby allowing the assessment of
metals long-term effects within resident communities at polluted
sites. Metal contamination may drive succession in algal communities
towards more pollution tolerant species (Gustavson and Wängberg,
1995), resulting in an increased tolerance of communities (Blanck
et al., 1988). It may also result in loss of species diversity (Leland and
Carter, 1984; Medley and Clements, 1998).
The purpose of this study was to:
• characterize the aquatic environmental conditions in the Coval da
Mó and Fílvida streams and determine the geochemical background
of trace metals in streambed sediment in order to evaluate the
degree of the dispersion of trace elements in streambed sediment.
• evaluate the trace-element concentrations found in surface water
and their potential detrimental effects on aquatic habitat in Coval da
Mó and Fílvida streams.
• compare communities growing at different sites along strong metal
pollution gradients.
2. The study area
2.1. Environmental setting
The Coval da Mó old mining area is located 20 km northeast of the
city of Aveiro in the drainage area of the Caima river (140 km2), a
tributary of the Vouga river (Fig. 1a). The mine is located in a moun-
tainous regionwith rounded rolling hills and short narrowvalleys that
form a dendritic-type drainage network. Coval da Mó and Fílvida
streams are small perennial impacted tributary streams with an ap-
proximate width of 1 m to 2.5 m, and 20 to 50 cm deep, respectively
(Fig. 1b, c).
Thisareahasaseasonaltemperateclimate,withaverywelldefined
rainyand dry season. Most rainfall occurs between October and March
and the dry season occurs from June to September (DGRAH, 1981).
According to “Instituto Nacional de Meteorologia e Geofísica” (INMG)
the long-term average annual precipitation is 1244 mm/year (pre-
cipitation range between 584 and 2514 mm/year — Borg and Hedlund,
2001). Temperatures range between 8.5 and 20.3 °C (annual mean of
13.7 °C).
Most of the landscape is forested (Eucalyptus sp. and Pinus pinaster
type), with some grassland to the south. Major human land uses in
this basin include agriculture, industry, mining, urban, and mixed use.
Agricultural activities include the production of fruits, grains and
vegetables.
2.2. Geological setting
Set in the Central Iberian Zone of the Iberian Massif, a tectonos-
tratigraphicunitoftheIberianPeninsulafirstdescribedbyRibeiroetal.
(1979), the mine is located in terrains of the Beira Schist Complex, a
meta-sedimentary formation mainly composed by phylites, meta-
greywacke and micaschists of still lower Cambrian age.
As mentioned before, the wall rock in the mine was mainly black,
chloritic schist from the Beira Schist Complex of Cambrian age, in
which faults and fracture lines of E–W, WNW–ESE and ENE–WSW
directions were filled by the mineralizing solutions of probable late to
post-Hercynian age (Thadeu, 1977).
2.3. Mining
The Coval da Mó mine is part of the mining complex of Braçal,
situated in the Aveiro district, of the Beira Litoral Province, Central-
West Portugal, about 30 km from Aveiro. Mining activities at this site
started in 1856, but due to economical difficulties caused by the First
World War period the mining complex was forced to close in 1918.
Afterwards, the second mining period started during the Second
World War in 1942, and lasted until the end of the 60's (total shut-
down in 1972).
The main ore exploited in this mine was lead in the form of galena
(PbS), and also some zinc [sphalerite — (Zn,Fe)S] and silver (in galena
and sphalerite) as accessories. Gangue minerals are dolomite (CaMg
(CO3)2),siderite(FeCO3),some(notmuch)quartz(SiO2),pyrite(FeS2),
somechalcopyrite(CuFeS2)andothersecondaryalterationmineralsas
aragonite (CaCO3) and anglesite (PbSO4). Macroscopic observations
permit to distinguish between three types of mineral association:
Type 1: is richer in sphalerite and pyrite with main gangue mineral
being quartz; Type 2: represents the main mineralized vein body
consistingofgalenainlargecubic-octahedralcrystalsandwhitesparry
dolomite; Type 3: consists of small veins and brecciation caused by
fracturesatnearperpendicularattitudetomainvein,mostlyfullyfilled
with massive sulphides (mainly galena or pyrite) and almost no con-
tent of gangue minerals (Marques de Sá, 2004, 2008).
Thetype ofwaste generated byminingactivitiesin theCoval daMó
mining areawas verylarge dumps of wall-rock and vein stones, debris
and tailings.
5621
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 3

Fig. 1. Maps showing (a) the drainage area of the Caima river (the shaded area indicates the drainage area of the Fílvida stream); Maps (b and c) showing sampling locations: C =
stream sediments, surface water and diatom samples.•= sediment samples to define the longitudinal gradient of metal pollution for the studied trace elements. (d) Photo showing
the location of tailing (T) and surface runoff material samples (SR).
5622
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 4

3. Materials and methods
3.1. Field sampled media and sampling methods
Tailings and samples representing material from surface runoff:
to investigate the impact caused by dismantling and erosion of the
tailings around the mine site, tailing samples (Tailing 1: T11to T12;
Tailing 2: T21to T24and Tailing 3: T31to T34— Fig. 1d) and also
material from surface runoff (SR1to SR3) were collected in the area.
Each tailing sample consisted of a mixture of sub-samples (composite
sample)representativeofeachlayerdistinguishedbyachangeincolour
while the surface runoff samples consisted of a mixture of sub-samples
taken within a 2×2 m area.
Stream sediment samples: for this study, stream sediments in the
neighbourhood of the Coval da Mó mine site were collected (Fig. 1c).
The sediment samples were taken along the Coval da Mó and Fílvida
streams. The establishment of the sampling locations along 7 km of a
severely impacted stream was based on the metal gradient previously
reported (Nunes et al., 2003). Streambed-sediment samples were
collected from39 sites, (Coval da Mó stream: C1 to C20; Fílvida stream
C21 to C39) in October 2001. Ten samples were collected from outside
the study area, in drainages that had a geologic setting similar to the
study area, to determine the geochemical baseline.
At each sampling site, streambed-sediment samples were col-
lected, usingaplasticscoop.Thesampleswerecomposited,wetsieved
with local stream water to pass through b2 mm plastic screen and
collected in a plastic gold pan. About 1 kg of material was sealed in
plastic containers at the site and transported back to the laboratory.
Surface waters: samples were collected at 6 selected sites: (a) in
Coval da Mó stream (sample C232 — in the vicinity of the Coval da Mó
mine and highly impacted by several metals and sample C79 — 2.5 km
downstream the mine and moderately impacted — see Fig. 2); (b) in
Fílvida stream (C85 — 6 km downstream the mine and weakly
impacted — see Fig. 2); (c) in Caima river (C93 — see Fig. 1a); (d) in
Vougariver(V14—seeFig.1a)andalsoin(e)theSerradaFreita(SF1—
representing a pristine area — see Fig.1a). Sampling was carried out in
July 2001 (summer), October 2001 (autumn) February 2002 (winter)
and May 2002 (spring). In order to assess the physical and chemical
characteristics of surface water, sampling was carried out on each
selected site (as close to the centre of the river as possible) in acid-
rinsedpolyethylenebottles(1 L).Watersampleswerekeptcoolat4 °C
in a cooled box prior to laboratory analyses. Samples collected for
dissolved metals (dissolved phase) were first filtered using a 0.45 μm
Milliporemembranefilters(ASTM,1984)thenwerepreservedinultra-
purenitricacid(samplesacidifiedtopHb2)topreventprecipitationof
metals and bacterial growth (dissolved phase). The filters were
preserved prior to analysis (particulate phase).
Diatom communities: epilithic diatom communities were sampled
accordingtoPrygielandCoste (2000), at thesame6 selected sampling
points as watersamples (C232, C79, C85, C93, V14 and SF1) during the
same referred seasonal periods. These sites are representative of the
observed metal gradient and also from the reference site (SF1) used
for comparison. Five removable boulders/pebbles were chosen from
a non-shaded zone, at water depth between 10 and 30 cm, carefully
avoiding stagnant waters. Diatom samples were obtained by scraping
theuppersurfaceoftheboulders/pebbleswithatoothbrush.Eachsam-
ple was split in two, one kept alive (without preservation) and the
otherpreservedwithformalinsolution(5%).Fromthefirstsub-sample,
an aliquot was cleaned using HNO3(65%) and potassium dichro-
mate (K2Cr2O7), at room temperature for 24 h, followed by several
Fig. 2. Distribution of Cd, Fe, Mn, Pb and Zn among the six chemical fractions studied in tailings T1 and T2 samples at the end of the Selective Chemical Extraction.
5623
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 5

centrifugations (1500 rpm) to wash the excess of acid. Samples were
then air-dried on a cover slip and mounted for permanent slides using
NAPHRAX®.
3.2. Analytical techniques
3.2.1. Chemical analyses
Tailings, material from surface runoff and stream sediment: forheavy
metal analysis samples were oven dried at a constant temperature
of 40 °C until a constant weight was attained before dry sieving.
The selected samples for Selective Chemical Extraction analysis (C232,
C79 and C85) were dried at room temperature in order to avoid the
overestimation of most mobile forms. These were disaggregated and
passed through a plastic sieve with a 177 μm aperture following the
method of Peacock et al. (1996). According to several authors (Salo-
mons,1980; Groot et al.,1982; Robbe,1984; Lucas et al.,1986; Rhoads
and Cahill, 1999; Swennen and Van der Sluys, 1998) heavy metals,
eitherofnaturaloranthropogenicorigin,accumulatemoreextensively
in the thinner fractions of sediments and, therefore, metal concentra-
tions decrease with the increase of the grain size. Bradshaw et al.
(1972) mentioned that the b170 μm provides a good compromise
between obtaining sufficient sample for analysis and provides a good
contrast between background and geochemical anomalies. Lottermo-
ser et al. (1999) suggest that the b170 μm is the most likely fraction to
reflect hydromorphically dispersed metals and metal pollution.
The b177 μm fraction was submitted to multi-elemental analysis
(ACME Anal. ISO 9002 Accredited Lab — Canada). A 0.5 g split was
leached in hot (95 °C) aqua regia (HCl–HNO3–H2O) for 1 h. After dilu-
tion to 10 ml with water, the solutions were analysed for 35 chemical
elements by Inductively Coupled Plasma-Atomic Emission Spectro-
metry (ICP-AES) for 32 elements. In this study emphasis is given to
cadmium (Cd), cobalt (Co), copper (Cu), iron (Fe), lead (Pb), manga-
nese (Mn), nickel (Ni), and zinc (Zn). The precision of the analytical
methods was based on the routine replicate sampling. The quality
of all analytical procedures was checked by incorporating random
duplicate samples and C3 and G-2 standards in each analytical set
(providedbytheACMEAnalyticalLaboratory).Theresultswerewithin
the 95% confidence limits of the recommended values for these cer-
tified materials. Overall analytical precision was ±5% for the heavy
metals.
Surface waters: the chemical analysis of water samples was carried
out using Inductively Coupled Plasma-Mass Spectrometry (ICP-MS)
for cations (Ca, Cd, Co, Cu, Fe, K, Mg, Na, Pb and Zn), Ion Chromato-
graphy for Cl−, NO3
metry for NO2
Grasshoff's,1976; Chaussepied,1977). Filters were dried and analysed
for Cd, Co, Cu, Fe, Pb and Zn by Atomic Absorption Spectrophotometry
after acid decomposition (HCl+HNO3+HF). Other water data col-
lectedandrecorded on-site,usinga multiprobeWTWMultilineP4 SET,
included pH, electrical conductivity (EC — expressed as μS cm−1at
20 °C), dissolved oxygen (DO) and percentage saturation values. To
measure Chemical Oxygen Demand (COD), samples were oxidized
using K2Cr2O7at 150 °C for 2 h. After reaching room temperature
the samples were analysed at 420 nm using a spectrophotometer
HACH DR2000 (Jirka and Carter,1975). A rigorous quality control pro-
gramwasimplemented,duringwaterchemicalanalysiswhichincluded
reagent blanks, duplicate samples, and certified reference materials
(STANDARD WASTWATRA6). Precision and bias of the chemical anal-
ysis were less than 10%.
Selective Chemical Extraction (SCE): in order to elucidate the mode
of occurrence of the metals in the tailing materials and stream
sediment samples, and establish the geochemical patterns of trace
metals, which are useful for predicting their release ability into the
aquatic environment and into the ecosystem, SCE technique was used.
Samples C232, C79 and C85 were sequentially treated with different
reagents so that metals with different affinities for the mineral matrix
−and SO4
2−(Gjerde,1986) and spectrophotocolori-
−(USEPA,1983) and NH4
+(Strickland and Parsons,1972;
could be released (Tessier et al., 1979; Meguellati et al., 1983; Rapin
andForstner,1983;Quevauvilleretal.,1994;Gómez-Arizaetal.,2000).
The scheme used follows a 6-step sequential extraction procedure,
proposed by Cardoso Fonseca and Ferreira da Silva (1998).
According to Cardoso Fonseca (1982), the following extractants and
operationally-defined chemical fractions were taken: a) Step 1: ammo-
nium acetate (1 M NH4Ac, pH 4.5) — water soluble and dissolved ex-
changeable ions, specifically adsorbed and carbonate-bound; b) Step 2:
hydroxylamine hydrochloride (0.1 M NH4OH.HCl, pH 2) — ions bound
in Mn oxyhydroxides; c) Step 3: ammonium oxalate (dark) (0.175 M
(NH4)2C2O4–0.1 M H2C2O4, pH 3.3 — Tamm reagent in dark) — ions
linked to amorphous Fe oxides; d) Step 4: H2O235% — ions associated
to organic matter (in this step sulphide-bound as primary sulphide
minerals could not be totally leached out according to Rapin and
Forstner(1983)andKheboianandBauer(1987));e)Step5:ammonium
oxalate (UV) (0.175 M (NH4)2C2O4–0.1 M H2C2O4, pH 3.3 — Tamm re-
agent under UV radiation — ions associated to crystalline Fe oxides; f)
Step6:mixed-acid(HCl+HNO3+HF)decomposition—ionsassociated
to matrix elements in lattice positions, resistant oxides and some sul-
phides. After each reaction timing, the solutions were centrifuged and
filtered. For metal speciation (Cd, Fe, Mn, Pb, and Zn) the sampled solu-
tions were analysed by Atomic Absorption Spectrometry (AAS) using
a GBC906 spectrophotometer. The accuracy of the sequential treat-
ment, considered as a whole, may be estimated by the comparison of
the total sum of the amounts obtained after each sequential extraction
step with the total amount obtained after hot mixed-acid attack of the
samesample(Cd:70–88%,Fe:74–84%,Pb:64–81%,Mn:74–79%,andZn:
58–72%).
3.2.2. Mineralogical analysis
Methods used to analyse the mineralogy of the selected samples
were:microscopictransmissionandreflectionstudy;microprobe(WDS)
study. Microscopic study involved both thin section observation of
wall-rock and transparent gangue minerals, as well as observation
of polished sections of sulphides and other opaque minerals. Micro-
probe analysis was conducted in INETI laboratory in 2004 (Marques
de Sá, 2004; Marques de Sá et al., 2005), in polished sections pre-
pared from samples collected at the dumps. The b177 µm fraction of
the stream sediment samples was mineralogically characterized by
powder X-ray diffraction (XRD) using a Philips X'Pert MPD, equipped
with an automatic divergence slit, CuKα radiation (20 mA and 40 kV)
and a Ni filter. For routine XRD inspections (for samples of stream
sediments and efflorescent salts) 4–70° 2θ scans were used with 0.5
counting time per step.
Diatom communities: for Scanning Electron Microscopy (JEOL-JSM
5400), a drop of the oxidized sample was placed on a metal stub
previously covered with a thin pellicle of carbon. The sample was
allowed to dry at room temperature. The stub with the sample was
thencoatedwithamixtureofgold-palladium.Diatomswereidentified
and quantified under the light microscope (Leitz Biomed 20EB) using
a 100× objective (N.A.1.32). A total of about 400 valves were counted
in each sample. Taxonomy was mainly based on Germain (1981),
Krammer and Lange-Bertalot (1986, 1988, 1991a, 1991b) and Prygiel
and Coste (2000).
3.3. Data analysis
Potential for adverse biological effects: various measures of biologic
responsetoelevatedconcentrationsofdeposit-relatedtraceelementsin
sediment are proposed by several authors and summarized in several
papers(MacDonaldetal.,2000;Ingersolletal.,2001).Theestimationof
potentialtoxicityofthecontaminantsinsedimentsampleswasbasedon
the Consensus-Based Sediment-Quality Guidelines (CBSQ)proposed by
MacDonald et al. (2000). The Threshold Effect Concentration (TEC) is
definedasthe“contaminantconcentrationbelowwhichharmfuleffects
on sediment-dwelling organisms were not expected,” whereas the
5624
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 6

Probable Effect Concentration (PEC) is defined as the “contaminant
concentration above which harmful effects on sediment-dwelling
organismswereexpectedtooccurfrequently”(MacDonaldetal.,2000).
Theseguidelinescanbeappliednotonlytoindividualcontaminants,
but also to a suite of metals and compounds in order to estimate their
combined toxic effects. The environmental-sample concentration is
divided by the Probable Effect Concentration (PEC), the concentration
above which, according to the Consensus-Based Sediment-Quality
Guidelines, adverse effects on benthic organisms are expected; the
resulting value is called a PEC quotient (PEC-Q). In the case of metals,
mean PEC-Qmetalsneeds to be calculated based on the summing of the
PEC-Q for the individual metals and dividing by the number of metals
involved. Finally, the mean PEC-Q metals are compared to values
associated with sediments of known toxicity to such standard test
organisms as the amphipod Hyalella azteca, or the insect larva Chirono-
mussp.Thus, this approach offers ameans ofestimatingthetoxicityof a
given sediment sample without actually performing the toxicity tests.
The Maximum Probable Background Concentration (MPBC) for a
metal at a reference site was calculated according to the method
proposed by CSST (2003) and using 10 sediment samples collected in
adjacent uncontaminated streams.
Diatoms: the methodology used for the analysis of species data was
Canonical Correspondence Analysis (CCA), which was used to study
diatom correlations with environmental variables. The main purpose of
thismethodistotreatthedataasawhole,sothatthediatomcommunity
structureisevidencedandtheinformationiscondensedinasimpleand
interpretable way. CCA was performed using the computer program
Canoco (version 4.5) (Ter Braak and Šmilauer, 2002). The matrix used
for CCA of surficial water samples is composed of 62 diatom taxa
(averagevalvenumberofthe4seasons)and12environmentalvariables
while for sediments the matrix is composed of 62 diatom taxa (average
valve number of the 4 seasons) and 9 environmental variables. Species
dataweresquareroottransformed.Oneoftheoptionsofthisprogramis
theprogressive selection of environmentalvariables,whichchoosesthe
setofvariablesthatbetterexplainsthespeciesdispersion.Thestatistical
meaning of each variable is tested with a Monte Carlo permutation test.
Thepowerofthistestincreaseswiththenumberofpermutations,sothe
maximum number possible (999) in the program was chosen. Only
significant variables (P≤0.05) were included in the analysis (Ter Braak
and Šmilauer, 2002).
Changes in diversity were checked by calculation of Shannon–
Weaver (H′) Index whichis widely used(Almeida,1998; Butcheretal.,
2003; Reiss and Kröncke, 2005) and was defined by Shannon (1948)
(in Washington, 1984):
H′= ∑
i=1
specimens of i specie; N — total number of specimens.
s
ni
Nlog2
ni
Nwhere: s — number of species; ni—number of
4. Results
4.1. Tailings and surface runoff materials
The tailing materials were physically and chemically heteroge-
neous. Tailings were composed by sand-silt sized materials that have
undergone partial oxidation since dumping.
According to the results of Table 1 the most abundant metals in the
tailings were Fe (3.80–20.1%), Pb (0.60–3.09%), Zn (481–3331 mg kg−1)
andNi(17–465 mgkg−1)followedbyCd(0.2–4.3 mgkg−1).Manganese,
Cu, and Co showed only distinct samples with high concentrations. These
results are probably best explained by the mineralogy of the deposited
tailings. Surface runoff materials showalsohigh contentsin Fe(6.1–7.3%),
Pb (0.66–1.75%), Zn (884–2294 mg kg−1) and Cd (2.9–8.1 mg kg−1).
The mineral suite in the tailing samples comprises the main
minerals present in the parental granite, which are ubiquitous: quartz,
alkali feldspar (NaAlSi3O8), biotite [K(Mg, Fe)3AlSi3O10(F,OH)2], and
muscovite [KAl2(AlSi3O10)(F,OH)2], or [(KF)2(Al2O3)3(SiO2)6(H2O)].
Mineralogicalresults frommacroscopic observation to microscope,
electron microprobe (WDS), XRD, cathodoluminescence and other
techniques of the ore minerals indicate that galena is the most
common sulphide and presents itself frequently as cubic-octahedral
crystals, mainly in Type 2 association. According to the description of
Marques de Sá (2008) it also occurs as thin films (Type 1) or forming
massive veins with cubic cleavage (Type 3). Minor elements were
studied with the electron microprobe (WDS) from LNEG laboratories
and revealed that galena was generally richer in Ag (mean value of Ag
is 0.08wt.% — Table 2).
Other minor components as Zn and Fe are thought to be absorbed
from contiguous or previous minerals. Argentiferous galena occurred
according to Cabral (1859) and Santos (1948) mainly in the massive
galena veins close to the surface.
Sphalerite (ZnS) is not as abundant as galena and occurs in two
different forms: sphalerite I (associated to Type 1) occurs as fibrous
spherical aggregates of dark brown colour, which previous authors
(Beutell and Matzke,1915; Jesus et al.,1930) identified as wurtzite [(Zn,
Fe)S].XRDdatapublishedbyMarquesdeSáandNoronha(2007) donot
support the statements of these authors. It is distinguishable from the
other sphalerite not only by its habit and colour but also on the basis of
its Fe content, which in sphalerite I is always above 5 wt.%. Sphalerite II
occursasisolatedsubhedralcrystalsormassesofdarkredcolourwithin
sacharoidal white dolomite and is generally associated with Type 2
association. Sphalerite II has an Fe content always below 5 wt.% and is
generallyricherinCd(meanvalue0.58 wt.%)thansphaleriteI,ascanbe
seen in Table 2.
Pyrite (FeS2) is also very common in this mine occurring in two
different genetic species: pyrite I is relatively abundant in Type 1
association and forms cubic, sometimes corroded crystals; pyrite II
occursinType2association,incubiccrystalsorsometimesincomposed
pyritohedral habit. Generally microprobe analysis of pyrite does not
show large differences albeit pyrite I is slightly more ferriferous than
pyrite II. Minor elements like Sb are more dominant in pyrite II, and
minor elements like Ni and Sn are more dominant in pyrite I.
Pyrrhotite (Fe1−xS2) occurs in stage one of the paragenetic
sequence but is generally embedded in galena crystals and masses in
the form of small (20–60 µm) pinkish crystals (observed in reflected
light microscopy). Marcassite (FeS2) and melnicovite (FeS2gel)-pyrite
are observed as late supergenic alteration minerals.
Chalcopyrite (CuFeS2) is rare and its most common occurrences
are late in the paragenetic sequence accompanying other alteration
minerals like anglesite (PbSO4).
The carbonate suite of minerals present includes dolomite CaMg
(CO3)2, ankerite CaFe(CO3)2, siderite (FeCO3) and calcite (CaCO3). The
Table 1
Concentration of Cd, Fe, Mn, Ni, Pb and Zn in tailings and surface runoff materials at
Coval da Mó mine (all values in mg kg−1except for Fe and Pb %).
SampleCdCo CuFeMnNi Pb Zn
Tailing 1T11
T12
Average
T21
T22
T23
T24
Average
T31
T32
T33
T34
Average
SR1
SR2
SR3
Average
4.3
0.2
2.3
1.3
0.2
0.7
9.6
3.0
3.6
3.2
3.6
2.4
3.2
8.1
3.9
2.9
5.0
33
4
19
17
6
5
123
38
492
267
380
257
247
223
314
260
389
174
250
180
248
156
168
113
146
10.90
9.90
10.40
10.50
20.10
7.20
8.50
11.58
14.50
3.80
13.60
4.60
9.13
7.30
7.20
6.10
6.87
843
110
477
301
37
107
1289
434
87
127
97
118
108
685
495
473
551
388
29
209
77
40
29
465
153
29
17
17
21
21
51
78
60
63
2.54
2.17
2.27
2.63
3.10
2.78
0.99
2.37
1.83
0.61
1.67
0.75
1.21
1.75
1.34
0.66
1.25
2269
646
1458
630
3331
309
2368
1660
2397
481
1962
456
1324
2294
1315
884
1498
Tailing 2
Tailing 3
4
3
2
3
3
Material from
surface runoff
37
21
18
25
5625
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 7

most abundant and main gangue mineral is dolomite. Three different
compositions occur: a first one in the centre of crystals with ankeritic
composition (Mg/Fe N4); a second with ferriferous dolomite composi-
tion with high Fe content (~4 wt.%); a third “pure” dolomite composi-
tionwith Mg/Fe bb4. It can show different colours dependingon Fe, Mg
and Mn content varying from translucid white to pinkish red or yellow.
Siderite occurs in Type 1 association, in the first paragenetic stage
and also shows zonation varying with Mg and Fe content. It occurs as
thin dark red films and masses covering or embedded in comb quartz.
Calciteisgenerallyalatestagemineralwhichfillscavitiesandfractures
and isn't very abundant.
Late stage supergenic alteration and oxidation minerals also include
anglesite, which fills fractures in galena, massicot (PbO) occurring as
yellowpowdersonoxidizedgalena,aragoniteasanalterationmineralof
carbonates in thin white crystals, and various Fe hydroxides.
Fig. 2 shows the results of Selective Chemical Extraction in selected
samples of tailing materials (T1 and T2 samples).
The results indicated that Pb is associated in several different
bearing phases, but mostly in the residual fraction as PbS [high
percentage of extraction associated to step 4 (7.4 to 11.1%) and 6 (31.2
to 57.4%) and in the exchangeable fraction, probably as PbCO3and
PbSO4(high percentage of extraction associated to step 1 — 16 to
48.4%). Some Pb was found in the oxide/hydroxide fraction. All the
values extracted by each reagent exceeded the PEC concentration
values proposed for Pb by MacDonald et al. (2000).
Also Zn and Cd were found in several different bearing phases, but
primarily in exchangeable fraction. The Zn showed a wide variation
between samples with high levels in the carbonate, oxide and residual
fractions. According to the results the exchangeable fraction accounts
forabout 47–73% of Cd adsorption, but its rolebecomes reduced for Zn
(27–39%).
Consequently, the residue has variable concentration fractions
depending on the metal considered, namely circa 0% for Cd, 31–57%
for Pb and 5–22% for Zn.
Table 2
Selected geochemical data from sulphides in Coval da Mó (values in wt.%).
AgBi Cd CuFe Mn NiPbS Sn Zn
Galena (n=45)Mean
Min
Max
SD
Mean
Min
Max
SD
Mean
Min
Max
SD
Mean
Min
Max
SD
Mean
Min
Max
SD
Mean
Min
Max
SD
0.08
b0.01
0.92
0.21
0.03
b0.01
0.12
0.05
0.03
b0.01
0.10
0.03
0.06
b0.01
0.42
0.12
0.09
b0.01
0.25
0.09
0.07
b0.01
0.25
0.09
–
–
–
–
–
–
–
–
0.05
0
0.53
0.13
0.01
b0.01
0.09
0.03
0.04
b0.01
0.17
0.06
0.10
b0.01
0.68
0.19
0.33
0.05
0.67
0.19
0.58
0.23
0.83
0.18
0.06
b0.01
0.31
0.09
0.04
b0.01
0.16
0.05
33.63
31.93
38.45
2.04
0.05
b0.01
0.17
0.06
0.10
b0.01
0.30
0.13
0.16
b0.01
0.33
0.13
0.04
b0.01
0.23
0.07
46.41
43.64
48.52
1.35
30.09
28.79
31.19
0.89
59.80
57.60
62.22
1.55
6.12
5.03
8.06
1.03
3.22
2.94
3.50
0.17
0.02
0
0.19
0.05
b0.01
b0.01
0.04
0.01
0.03
b0.01
0.06
0.03
0.07
b0.01
0.67
0.18
0.03
b0.01
0.05
0.02
0.02
b0.01
0.09
0.03
0.04
b0.01
0.20
0.06
0.52
b0.01
3.08
0.82
0.04
b0.01
0.10
0.04
0.31
b0.01
0.51
0.14
0.02
b0.01
0.07
0.03
0.04
b0.01
0.12
0.05
85.56
83.29
87.67
13.59
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
13.57
12.96
14.18
2.16
52.18
49.23
53.66
1.33
34.64
33.86
35.63
0.55
38.65
36.69
40.82
1.20
33.71
32.48
34.85
0.86
33.47
32.42
34.19
0.52
–
–
–
–
0.01
0
0.11
0.03
0.70
0
1.93
0.68
–
–
–
–
–
–
–
–
–
–
–
–
0.22
b0.01
3.18
0.82
0.03
b0.01
0.11
0.04
0.18
b0.01
1.17
0.41
0.07
b0.01
0.27
0.09
58.11
56.43
62.05
2.03
60.84
59.63
62.24
0.84
Pyrite (n=14)
Chalcopyrite (n=8) 0.01
b0.01
0.05
0.02
0.03
b0.01
0.31
0.09
0.02
b0.01
0.09
0.03
0.05
b0.01
0.32
0.11
Phyrrhotite (n=13)
Sphalerite I (n=7)
Sphalerite II (n=9)
Table 3
Mean, median and range of the Cd, Co, Cu, Fe, Pb, Mn, Ni and Zn concentrations in sediment samples and reference values available in the literature.
Cd CoCuFe PbMnNiZn
Coval da MóMean
Median
Range
Mean
Median
Range
Mean
Median
Range
Mean
Median
Range
Mean
Mean
Mean
Mean
Mean
Mean
5.8
6.6
0.0–10.8
1.8
1.8
0.9–2.6
0.6
0.4
0.2–1.6
0.2
0.1
0.1–0.5
0.3
b0.2
b0.2
0.1
1.0
0.2
34
36
21–39
30
30
27–34
7
7
4–15
5
3
1–30
9
5
16
24
30
32
75
65
34–115
39
38
33–49
35
30
15–70
13
11
3–48
32
27
36
25
n.a.
68
4.44
4.29
3.30–6.13
2.89
2.88
2.54–3.23
2.02
2.05
1.41–2.78
1.71
1.71
0.46–3.48
2.48
1.75
3.42
4.32
n.a
3.21
5875
5463
693–23,920
876
876
513–1391
185
58
30–1870
16
16
2–33
30
24
36
14.8
70
77
332
384
55–603
291
289
248–358
7
7
4–15
139
99
12–623
467
462
891
716
n.a
260
69
70
30–109
43
42
37–49
14
14
8–19
9
8
1–23
21
8.0
32
56
70
41
1020
1183
88–2135
395
390
274–487
122
121
70–212
53
47
11–125
111
98
84
65
175
159
Fílvida
Vouga(a)
Serra Freita(a)
ZCI1
GR2
RM3
CC4
SED5
MPBC6
ZCI — Central Iberian Zone; RG — Granitic rocks; MR — Metamorphic rocks; CC — Continental Crust; SED — European USA and Canadian lake sediments; FG — Geochemical background
estimated according to the methodology proposed by Lepeltier (1969).1, 2, 3 (data from Ferreira, 2000 in Nunes, 2007); 4 (data from Wedelpohl,1995); 5 (data from Hakanson,1980);
MPBC — Maximum Probable Background Concentration for a given metal, estimated according the methodology proposed by MacDonald et al. (2000); (a) data from Nunes (2007).
5626
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 8

Fig. 3. Plot of the trace metals concentration versus distance from the Coval da Mó mine.
5627
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 9

4.1.1. Stream sediment samples
Theresultsofthestreambed-sedimentfocusontheconcentrationsof
eight metals: Cd, Co, Cu, Fe, Pb, Mn, Ni, and Zn from among the 32
elements analysed. These elements were selected based on the ACP
results(Nunes,2007)whichallowedtheidentificationoftheseelements
as anthropogenic association. Rock-forming trace elements, such as Al,
Ba,Ga,K,Mg,Na,Sc,TiandVshowednoabruptchangesinconcentration
in the streambed-sediment downstream from the Coval da Mó mine.
Table3showstheresultsofthemeanvaluesobtainedinCovaldaMóand
Fílvida streams with the mean values proposed by different authors.
The results show that the main drainage system affected by the
abandoned mine is the Coval da Mó stream, which joins Fílvida stream
about 2.5 km downstream, which in turn joins the Caima river about
3 km downstream (Fig. 1c). According to the geochemical results
(Table 3), Cd, Pb, Mn, Ni and Zn concentrations in the stream sediments
differed widely, but the highest concentrations were generally detected
in samples collected in Coval da Mó stream (Pb: 693–23,920 mg kg−1;
Cd: 0.0–10.8 mg kg−1; Zn: 88–2135 mg kg−1). Samples from the Fílvida
riverhavelowerconcentrationsthanthosefromCovaldaMóstream(Cd:
0.9–2.6 mg kg−1; Pb: 513–1391 mg kg−1and Zn: 274–487 mg kg−1).
Vouga river shows a substantial decrease in the studied elements con-
centrations. According to the XRD results, the streambed-sediment sam-
ples show the presence of quartz, feldspar, biotite, muscovite, ilite and
kaolinite. Other minerals as pyrite, galena, sphalerite, anglesite and
cerussite also occur in the stream sediments collected near the Coval
da Mó mine.
Fig. 3 shows the metal pollution gradient in Coval da Mó and
Fílvida streams estimated for Cd, Fe, Pb, Mn, Ni and Zn.
Theincreaseofdeposit-relatedtrace-elementconcentrationsinthe
streambed sediment is about two orders of magnitude comparatively
Fig. 4. Distribution of Cd, Fe, Mn, Pb and Zn among the six chemical fractions studied in C232, C79 and C85 samples at the end of the Selective Chemical Extraction.
5628
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 10

to estimated maximum probable background values and by one order
of magnitude for Zn and Cd (Table 3). The concentrations in the Coval
daMóandFílvidastreamswere alsohigherthan thevaluesfor granitic
and metamorphic rocks. The obtained values for Coval da Mó also
exceeded the values proposed for Central Iberian Zone (ZCI). Direct
field evidence showed transport of tailing materials from the Coval da
Mó mine site into the Coval da Mó stream. Especially close to the mine
this is clearly seen (sample C232) which has almost the same concen-
tration of metals as the tailing samples.
The PEC for Cd (5 mg/kg) is greater than the Ecological Screening
Value for Sediment (ESV — 1.0 mg/kg) or the Screening Level Concen-
tration (SLC — the highest concentration of a contaminant that can
be tolerated by approximately 95% of the benthic fauna — Cd: 0.6–
1.0 mg kg−1). Seventy seven percent of total samples exceeded the
PEC values and all the samples have concentrations higher than the
ESV and SLC values; noneof these values exceeded the human-contact
criterion of 30 mg kg−1(Breault et al., 2000; Zimmerman and Breault,
2003).
For Fe 41% exceeded slightly the PEC values (samples CM5 to
sample CM20). The ratio values between the concentrations and the
PEC value ranged from 0.6 to 1.5 (Coval da Mó stream: 0.8–1.5; Fílvida
stream: 0.6–0.8). The same samples, mainly in the Coval da Mó
stream, showed Ni concentrations that generally exceeded the PEC
value (ratio values ranging from 1.2 to 2.2) but none of these values
exceeded the human-contact criterion of 300 mg kg−1(Breault et al.,
2000; Zimmerman and Breault, 2003).
All of the Pb concentrations exceeded the TEC and PEC values. The
ratio between the concentration and the PEC value ranged from 3.9 to
184 (Coval da Mó stream: 5.3–184; Fílvida stream: 3.9–10.7). Also all
the samples exceeded the Ecological Screening Value for Sediment
(30.2 mg kg−1) or the Screening Level Concentration (the highest
concentrationofacontaminantthatcanbetoleratedbyapproximately
95% of the benthic fauna — Pb: 30–47 mg kg−1). All the samples
exceeded also the human-contact criterion of 300 mg kg−1(Breault
et al., 2000; Zimmerman and Breault, 2003).
For Mn, one sample exceeded the TEC value (460 mg kg−1) but
none exceeded the PEC value. The ratio between the concentration
and the PEC value ranges from 0.1 and 0.6 (Coval da Mó stream: 0.1–
0.6; Fílvida stream: 0.2–0.3).
Thirty percent of the samples had Zn concentrations exceeding the
PECandtheratiovaluesbetween0.1and4.6(CovaldaMóstream:0.1–
4.6; Fílvida stream: 0.6–1.1). Only four samples collected upstream the
mine site did not exceed the Ecological Screening Value for Sediment
(124 mg kg−1) or the Screening Level Concentration (the highest con-
centration of a contaminant that can be tolerated by approximately
95% of the benthic fauna — Zn: 120–160 mg kg−1). None of the de-
tected Zn concentrations exceeded the human-contact threshold of
2500 mg kg−1(Breault et al., 2000; Zimmerman and Breault, 2003).
In order to complement the results obtained with ICP-MS, a 6-step
sequential chemical extraction (SCE) was carriedouton three samples
(C232, C79 and C85) from the impacted streams. The same samples
were also analysed by XRD for the determination of mineralogical
composition. Experimental results of the sequential extraction steps
in samples of streambed sediments are provided in Fig. 4, and it can be
noted that a significant percentage of metals were actually adsorbed
to the mineral surfaces, some as surface complexes and others as sul-
phide minerals.
According to the results, the largest proportion of Pb is associated
withtheexchangeablefraction(Step1)(percentageofextractionranges
between 44 and 71%). Also the sulphides/organic matter (Step 4) and
the residual forms (Step 6) are important. All the Pb concentrations
obtained in each step exceeded the PEC value.
Zinc is extracted mainly by H2O2reagent (percentage of extraction
ranges between 41 and 50%) and by ammonium acetate. In the case of
Cd, the results are similar to Zn. The exchangeable Cd concentration is
also important (18–46% of extraction).
Table 4Physicochemical characteristics of water at C232, C79 C85, C93, V14 e SF sites.
Summer
Autumn
Winter
Spring
C232
C79
C85
C93
V14
SF
C232
C79
C85
C93
V14
SF
C232
C79
C85
C93
V14
SF
C232
C79
C85
C93
V14
SF
pH
6.5
7.2
7.2
7.1
7.4
7.2
5.7
7.5
7.0
7.0
7.1
7.7
6.5
7.4
7.5
7.5
7.0
7.6
6.1
6.8
6.7
6.7
7.1
6.2
Cond
μS cm−1
200
130
96
104
81
33
144
135
93
74
82
19
147
151
103
100
104
24.5
212
155
90
101
75
18
O2
mg L−1
4.8
8.3
9.0
9.0
9.9
10.4
8.4
9.8
10.2
10.2
9.4
10.8
9.5
11.1
11.2
11.3
11.9
11.3
9.1
9.1
10.2
9.6
9.9
10.0
O2
% sat.
50
90
98
90
117
110
85.0
96.0
100.0
96.4
98.0
97.7
88.5
99.5
102.8
101.5
105.7
104.3
101.8
93
94
102
102.7
100.2
COD
mg L−1
bdl
9.3
0.0
5.5
6.3
5.0
98.5
4.5
4.7
2.7
5.0
4.5
9.7
5.3
8.3
8.0
11.5
bdl
9.3
10.3
13.7
16.7
bdl
12.3
Cl−
mg L−1
7
10
8
13
11
4
7
9
7
9
8
3
6
11
9
9
8
2
7
12
8
10
9
3
SO42−
mg L−1
40
17
13
7
6
1
74
8
13
9
7
2
36
26
8
5
5
bdl
5
30
16
6
4
1
NO3
−
mg L−1
bdl
2
4
8
5
1
2
6
4
6
6
1
bdl
2
6
7
5
1
bdl
1
4
6
5
bdl
NO2
−
μg L−1
bdl
bdl
bdl
bdl
31
bdl
11
64
12
14
28
10
17
19
22
50
67
20
26
19
26
149
51
24
NH4
+
mg L−1
0.03
0.03
0.04
0.04
0.13
0.05
0.03
0.05
0.04
0.06
0.02
0.03
0.02
0.03
0.02
0.07
0.06
0.03
0.02
0.03
0.02
0.10
0.04
0.03
Cd (d)
μg L−1
bdl
2
2
bdl
bdl
bdl
10
bdl
bdl
bdl
bdl
bdl
5
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
Fe (d)
μg L−1
bdl
bdl
bdl
bdl
bdl
bdl
20
70
90
80
70
30
80
100
20
20
170
20
100
20
20
120
90
30
Co (d)
μg L−1
9
bdl
bdl
bdl
bdl
bdl
14
bdl
bdl
bdl
bdl
bdl
10
bdl
bdl
bdl
bdl
bdl
9
bdl
bdl
bdl
bdl
3
Cu (d)
μg L−1
bdl
6
bdl
bdl
bdl
bdl
10
4
4
4
4
bdl
2
3
2
32
3
bdl
5
11
14
8
5
bdl
Pb (d)
μg L−1
bdl
bdl
bdl
bdl
bdl
bdl
1734
bdl
bdl
bdl
bdl
bdl
43
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
Zn (d)
μg L−1
177
74
5
9
3
3
1127
16
34
17
27
6
695
9
12
58
16
13
157
87
114
14
7
6
Cd (p)
μg L−1
48
6
7
8
9
7
bdl
3
4
19
5
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
Fe (p)
μg L−1
7690
bdl
bdl
103
184
72
126
218
135
116
230
bdl
1347
bdl
149
174
bdl
bdl
4020
bdl
bdl
213
311
bdl
Co (p)
μg L−1
bdl
bdl
bdl
bdl
bdl
bdl
bdl
bdl
3
bdl
bdl
bdl
bdl
bdl
bdl
bdl
3
bdl
bdl
bdl
bdl
bdl
bdl
bdl
Cu (p)
μg L−1
bdl
bdl
bdl
bdl
bdl
bdl
bdl
14
17
13
29
bdl
bdl
bdl
bdl
bdl
37
bdl
bdl
2
bdl
2
3
bdl
Pb (p)
μg L−1
bdl
bdl
22
8
17
bdl
34
bdl
bdl
26
bdl
bdl
508
19
bdl
13
13
bdl
59
bdl
bdl
bdl
bdl
bdl
Zn (p)
μg L−1
74
bdl
bdl
bdl
bdl
bdl
9
9
10
12
bdl
3
36
196
35
18
12
124
191
10
11
14
18
13
% sat — % of saturation; bdl — below detection limit (Cd and Cu: 2 μg L−1; Co and Zn: 3 μg L−1; Pb: 10 μg L−1; Fe: 20 μgL−1; SO42−and NO3
−: 0.1 mg L−1); (d) dissolved component; (p) particulate component.
5629
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 11

4.2. Surface water
The physical and chemical parameters determined for surface
water sites at different seasons (Table 4) show that the average pH is
around 7, with conductivity ranging from 90 to 212 µS cm−1, high
oxygen content, loworganic content (COD). It is possible to observe in
all the campaigns a slight pH decrease at C232 due to the influence of
Coval da Mó mine. In the studied area there are no cases of very high
nutrient content, although an increasing gradient of NO3
can be found from C232 (forest region) to C85 (urban and agricultural
area — NO2
ranging from 6 mg L−1to 12 mg L−1.
−and NO2
−
−: 22 g L−1). Concentrations of Cl−were low, with levels
Table 5
Relative abundance (%) and labels of diatom taxa at sites C79, C85, C93, C232, V14 and SF1 with indication of number of counted taxa and Shannon–Weaver's diversity index (H′).
Taxa LabelC79C85 C93C232 V14 SF1
Achnanthes oblongella Östrup
Achnanthes subhudsonis Hustedt
Achnanthidium minutissimum (Kütz.) Czarnecki
Achnanthidium minutissimum var. incospicua Oestrup
Achnanthidium minutissimum var. jackii (Rabenhorst) Lange-Bertalot
Achnanthidium saprophilum (Kobayasi et Mayama) Round & Bukhtiyarova
Anomoeoneis brachysira (Breb.) Grunow var. zellensis (Grunow) Krammer
Caloneis bacillum (Grunow) Cleve
Cocconeis placentula Ehrenberg var. placentula
Craticula submolesta (Hust,) Lange-Bertalot
Cyclotella meneghiniana Kützing
Diadesmis contenta (Grunow ex V. Heurck) Mann
Diatoma mesodon (Ehr.) Kützing
Encyonema minutum (Hilse in Rabh.) D. G. Mann
Encyonema silesiacum (Bleisch in Rabh.) D. G. Mann
Eolimna minima (Grunow) Lange-Bertalot
Eunotia bilunaris (Ehr.) Mills var. bilunaris
Eunotia bilunaris (Ehr.) Mills var. mucophila Lange-Bertalot Norpel & All
Eunotia exigua (Brebisson ex Kützing) Rabenhorst
Eunotia intermedia (Krasske ex Hustedt) Nörpel & Lange-Bertalot
Eunotia minor (Kütz.) Grunow in Van Heurck
Eunotia sudetica O. Muller
AOBG
ASHU
ADMI
AMII
AMJA
ADSA
ABZE
CBAC
CPLA
CSBM
CMEN
DCOT
DMES
ENMI
ESLE
EOMI
EBIL
EBMU
EEXI
EUIN
EMIN
ESUD
3.63 0.34
9.37
12.89
12.95
2.5
7.34 63.3331.99
0.39
78.690.89
0.67
1.340.35
0.13
0.43
1.69
0.33
0.37
0.35
0.47
6.38
6.79
3.21
10.69
1.75
43.13
2.17
6.73
15.162.171.21 0.83
26.52
0.76
0.95
0.89
0.64
0.26
2.66
2.760.39
Fistulifera saprophila (Lange-Bertalot & Bonik) Lange-Bertalot
Fragilaria biceps (Kütz.) Lange-Bertalot
Fragilaria capucina Desmazières var. capucina
Fragilaria capucina Desmazières var. rumpens (Kütz.) Lange-Bertalot
Fragilaria capucina Desmazières var. vaucheriae (Kütz.) Lange-Bertalot
Fragilaria cf. crotonensis Kitton
Fragilaria pinnata Ehrenberg var. pinnata (Staurosirella)
Fragilaria ulna (Nitzsch.) Lange-Bertalot var. ulna
Gomphonema parvulum Kützing
Luticola goeppertiana (Bleisch in Rabenhorst) D. G. Mann
Luticola mutica (Kützing) D. G. Mann
Mayamaea atomus var. permitis Hustedt Lange-Bertalot
Melosira varians Agardh
Navicula cryptocephala Kützing
Navicula gracilis Ehrenberg
Navicula gregaria Donkin
Navicula lanceolata (Agardh) Ehrenberg
Navicula leptostriata Jorgensen
Navicula rhynchocephala Kützing
Navicula veneta Kützing
Nitzschia brevissima Grunow
Nitzschia clausii Hantzsch
Nitzschia linearis (Agardh) W. M. Smith var. linearis
Nitzschia linearis (Agardh) W. M. Smith var. tenuis (W. Smith) Grunow
Nitzschia palea (Kütz.) W. Smith
Nitzschia perminuta (Grunow) M. Peragallo
Peronia fibula (Breb.ex Kütz.) Ross
Pinnularia divergens W. M. Smith var. divergens
Pinnularia microstauron (Ehr.) Cleve var. microstauron
Pinnularia subcapitata Gregory var. elongata Krammer
Pinnularia subcapitata Gregory var. subcapitata
Placoneis clementis (Grun.) Cox
Planothidium frequentissimum (Lange-Bertalot) Lange-Bertalot
Planothidium lanceolatum (Brebisson ex Kützing) Lange–Bertalot
Psammothidium bioretii (Germain) Bukhtiyarova et Round
Psammothidium helveticum (Hustedt) Bukhtiyarova et Round
Psammothidium subatomoides (Hustedt) Bukht. et Round
Reimeria sinuata (Gregory) Kociolek and Stoermer
Sellaphora seminulum (Grunow) D. G. Mann
Surirella roba Leclercq
Number of counted taxa
H′
FSAP
FBCP
FCAP
FCRU
FCVA
FCRO
FPIN
FULN
GPAR
LGOE
LMUT
MAPE
MVAR
NCRY
NGRA
NGRE
NLAN
NLST
NRHY
NVEN
NBRE
NCLA
NLIN
NZLT
NPAL
NIPM
PFIB
PDIV
PMIC
PSEL
PSCA
PCLT
PLFR
PTLA
PBIO
PHEL
PSAT
RSIN
SSEM
SRBA
8.96
14.69
0.59
7.03
1.57
0.06
12.471.49
1.75
0.74
2.05
0.349.69
0.06
0.56
1.29
1.62
1.34
0.60
1.01
1.68 0.401.91510.90 0.38
2.35
0.49
0.39
0.33
4.44
0.34
0.080.672.56
2.52
1.742.55
0.68
9.76
0.88
0.06
0.57
5.48
0.62
0.28
0.41 0.25
0.06
0.28
5.60
2.93
1.98 0.67
0.63
4.63
1.94
3.34
1.71
1.96
0.34
0.33
0.45
0.26
0.60
0.42
1.20
0.26
0.53
0.390.63
0.47
0.19
5.85 0.350.26
4.55
0.31
35.75
28
2.3
14
2.0
22
2.7
23
2.7
5
1.3
29
3.5
5630
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 12

Most of the dissolved and particulate trace metal concentrations
werelowor belowthe detection limit, withthe exception of Cd, Cu, Fe,
Pb and Zn which had higher concentration at site C232 near the Coval
da Mó abandoned mine.
4.3. Periphytic diatom communities
According to diatom results, important structural and composi-
tional differences were detected in the epilithic diatom communities
from the stream boulders/pebbles of the sampling points.
Approximately twenty different diatom taxa were counted at C79
(downstream of Coval da Mó mine) and C85 (downstream of Fílvida
stream) sites (Table 5).
The number of counted and identified taxa, was smaller at C79
revealing lower diversity (H′=2.0) than site C85 (H′=2.7). Similar
changeshavebeenreportedbyotherauthors,withdiatomdiversityand
species richness decreasing with increasing metal pollution (Say and
Whitton, 1980; Roch et al., 1985; Deniseger et al., 1986; Medley and
Clements, 1998; Dickman, 1998; Sabater, 2000). Abnormal valves of
Fragilaria capucina var. rumpens are represented in Fig. 5. At C79 Ach-
nanthidium minutissimum dominated, with more than 60% relative
abundance, only distantly followed by the deformed F. capucina var.
rumpens (ca.13%). Site C85 revealed a more stable community,
dominated by A. minutissimum closely followed by Fragilaria biceps,
F.capucinavar.rumpens,EncyonemaminutumandEncyonema silesiacum
with similar relative abundances (Table 5).
At site C93 (downstream of the Fílvida river mouth in river Caima)
the most abundant species were Eolimna minima and E. minutum and
H′ was 2.7.
At the mine site C232 only five taxa were counted (Table 5). The
most abundant ones were A. minutissimum (relative abundance of ca.
70%) and Gomphonema parvulum (relative abundance of ca. 11%).
In Vouga river E. minima and A. minutissimum were the most
abundant taxa (relative abundance of ca. 15%). The diversity index
value, at this site, was the highest (H′=3.5) which is concordant with
a stable and diverse community.
At the Serra da Freita Surirella roba (relative abundance of ca. 36%),
Eunotia bilunaris (relative abundance of ca. 27%) and Achnanthes
oblongella were the predominant taxa. The diversity was moderately
high (H′=2.7). This was chosen as the reference site, because it was
not subjected to anthropogenic influence.
Fig. 6a shows the CCA ordination of the first two axes for the
surficial waters, considering the environmental variables and sam-
pling points while Fig. 6b shows the plot of taxa versus environmental
variables.Thecumulativevariancepercentageofenvironment-species
data for each axis was 32.7% (axis 1), 56.1% (axis 2), 72.5% (axis 3),
87.0% (axis 4). Axes 1 and 2 explained about 90% (88.8%) of the
variance in the species data and in the environment-species relation.
The total inertia was 1.95.
Axis 1 shows negative correlationwith SiO2, conductivity, Cu and Fe.
Axis 2 reveals a strong positive correlation with SO4
NiNMnNCdNZn and a negative correlation with Fe.
2−NconductivityN
Fig. 5. Scanning electron microscopic photos (JEOL JSM-5400) of teratological valves of
FragilariacapucinaDesmazièresvar.rumpens(Kützing)Lange-Bertalot.(a)and(c)external
valve views; (b) internal valve view.
Fig. 6. CCA ordination of the first two axes showing scores for: (a) surficial water
samples and environmental variables. Each site is a letter (that corresponds to the name
of the river: C — Caima; SF — Serra Freita; V — Vouga) and a number (which represents
the sampling site) combination. Environmental variables are represented by vectors;
(b) for taxa and environmental variables. Taxa labels consist of a 4 letter code (for
further information see Table 5 and Appendix 1.1 to 1.4): the species closest to the tip of
the arrow of an environmental variable were the most correlated to it.
5631
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 13

According to Fig. 6a, sampling sites are well discriminated.
Quadrant 2 shows C232 and C79 (Caima river), the most impacted
sites by trace metals, followed by C93 and V14 in Quadrant 3. The
background sample, SF1, is clearly secluded in Quadrant 1. These
results point out that spatial variation is the most important factor.
The relationship between environmental variables and diatoms is
shown in Fig. 6b. For interpretation of diatom labels, see Table 5. This
diagram shows that all taxa in Quadrants 2 and 3 are correlated with
high concentrations of trace metals: Cd, Co, Cu, Fe, Ni, Mn, Pb, Zn and
SO4
laria biceps, A. minutissimum, Planothidium frequentissimum, Sella-
2−. Among these taxa we can find F. capucina var. rumpens, Fragi-
phora seminulum among others formerly referred. Diatom community
compositionwas shifted from less metal tolerant taxa to more tolerant
ones (from Quadrant 2 to Quadrant 3).
Fig. 7a showstheCCAordinationofthefirsttwoaxesconsideringthe
stream sediment sampling points and the environmental variables
whileFig. 7b shows theplotof taxaversus environmentalvariables.The
cumulative variance percentage of environment-species data for each
axiswas32.7%(axis1),56.1%(axis2),72.5%(axis3),87.0%(axis4).Axes
1 and 2 explained about 90% (88.8%) of the variance in the species data
and in the environment-species relation. The total inertia was 1.95.
Axis 1 had negative correlationwith Cu and Co. Axis 2 had a strong
positive correlation with Fe, Cd, Zn, Mn and Pb respectively and a
negative correlation with Ni.
Comparing Fig. 6 with Fig. 7 it's possible to see that both, surficial
waters and stream sediment data, show the same sites and taxa
distribution (in the same quadrants). As in surficial waters, Quadrants 2
and 3 in both figures are correlated with high concentrations of trace
metals.
5. Discussion of the results
In order to integrate the results from the stream sediments, surface
waters and periphytic diatom communities a Canonical Correspon-
dence Analysis was performed (Figs. 6 and 7).
CCA results showed that the studied stream systems were
influenced by the tailing deposits; (a) high concentrations of trace
elements (Cd, Fe, Pb, Ni and Zn) were found in the sediments of Coval
da Mó stream (downstream the mine). Copper, Pb, Zn, and Cd
concentrations in Coval da Mó stream sediments increased drastically
and directly below the mine (site C232). Also Mn follows the same
pattern but the contamination is not as high as for the other metals.;
(b) there was an increase in drainage waters acidity and trace metal
concentrations in surface water samples under influence of the tailing
material; (c) downstream from the confluence of Coval da Mó stream
with Fílvida river, deposit-related trace-element concentrations were
diluted by uncontaminated streambed sediment originating from the
upper part of Fílvida river (C79); diatom communities showed
different structure and taxonomic compositions along the trace
metal gradient; morphological deformities in diatom valves were
more numerous in metal polluted sites.
In the study area trace metal concentration decreased exponen-
tially with increasing distance to the tailings, almost reaching the
background levels 4 km downstream (C85 — Fig. 3).
Results showed that metal concentrations in stream sediment
samples exceeded criteria for benthic organisms (MacDonald et al.,
2000). In general, the samples with the highest metal concentrations
came from sites located in the Coval da Mó stream.
Experimental results of the sequential extraction steps in samples
of stream sediments show that a significant percentage of metals are
adsorbed to the mineral surfaces, some as surface complexes but also
insulphideminerals.Accordingtotheresults,thelargestproportionof
Pb is associated with the exchangeable fraction, sulphides and the
residual forms. The high levels of Pb related to the ammonium acetate
(11,860 mg kg−1) are connected to the occurrence of neoformed
minerals of lead such as cerussite and anglesite whereas the Pb
extracted by hydrogen peroxide (2113 mg kg−1) and acid decomposi-
tion (670 mg kg−1) is related to galena.
Because Pb is the most important anthropogenic element and
occurs as a sulphide phase (PbS), a small heterogeneous distribution
of Pb particles in sediment can easily appear as responsible for the
heterogeneity of the results. According to several authors (Laville-
Timsit and Wilhelm,1979; Wilhelm and Laville-Timsit,1979) sulphide
oxidation is responsible for the formation of stable secondary phases
that are the main vector of metals: oxyhydroxides of iron and also Pb
neoformed minerals such as anglesite. The formation of anglesite is
fast but this phase is unstable in the presence of carbonic acid and is
Fig. 7. CCA ordination of the first two axes showing scores for: (a) stream sediment
samplesandenvironmentalvariables.Eachsite isaletter(thatcorrespondstothe name of
the river: C — Caima; SF — Serra Freita; V — Vouga) and a number (which represents the
sampling site) combination. Environmental variables are represented by vectors; (b) for
taxa and environmental variables. Taxa labels consist of a 4 letter code (for further
information see Table 5 andAppendix1.1 to 1.4):the speciesclosesttothetipof the arrow
of an environmental variable were the most correlated to it.
5632
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 14

slowly substituted by the carbonate-bearing phase (cerussite). The
probable reactions are expressed by the following formulas:
PbS þ 2O2→PbSO4
and
ð1Þ
PbSO4þ H2CO3→ PbCO3þ H2SO4
Zn is extracted mainly by H2O2reagent and by ammonium acetate,
indicating that Zn may be associated with sphalerite and also with
smithsonite. Also amorphous and crystalline iron oxides (goethite
enriched in Zn) are important metal bearing phases. In the case of Cd,
the results are similar to Zn because sphalerite is enriched in Cd. The
exchangeable Cd concentration is important indicating that Cd may be
easily released and transported into the environment.
The presence of a downstream gradient concentration in metals
and the lower concentration of trace metals in the surface water
samples suggest that the dispersion is mainly mechanical (presence of
pyrite, galena and sphalerite in the stream samples) and that the
carbonate phases in the tailing materials act as an important factor to
neutralize the media and avoid the presence of acid mine drainage.
The taxa identified at C232 (Figs. 6b and 7b, Table 5) were A. minu-
tissimum (dominant taxon), Eunotia exigua is α-mesosaprobiont
according to Van Dam et al. (1994) and shows high tolerance to
contamination (Baffico et al., 2004) and to large spectra of chemical
pollutants (Guasch etal.,1998),G.parvulum,Cocconeisplacentula,andE.
minima. The main reason for the rareness of diatoms at this site was,
most probably, the high concentration of trace metals in the stream
sediments.
The results showed that A. minutissimum was the dominant
species at C79 too, where higher metal concentration in the sediments
(Pb, Cd and Zn) was found in comparison to C85. Gold et al. (2003)
hypothesize that the resistance of this taxon may be due to the
manner in which these adnate types with a short mucilaginous stalk
live below many other periphytic taxa. The periphyton protect them
with extracellular polymeric substances, thereby complexing metals
and reducing metal bioavailability and toxicity.
A. minutissimum a widely distributed taxon, has been frequently
pointed out, as an oligotrophic and oligosaprobic taxon indicating good
water quality (Slàdecek, 1986; Leclercq and Maquet, 1987; Prygiel and
Coste, 2000). Nevertheless, the resistance to metals of A. minutissimum
is still under discussion, showing contradictory results when literature
is consulted. Besch et al. (1972), Rushforth et al. (1981), Austin and
Deniseger (1985), Roch et al. (1985) and Sabater (2000) reported
decreasing abundances of this species in polluted sites. On the other
hand,Denisegeretal.(1986),Genteretal.(1987),MedleyandClements
(1998), Ivorra et al. (1999), Gold et al. (2002, 2003), and Cattaneo et al.
(2004) showed that A. minutissimum developed abundant populations
at sites contaminated with different metals (Pb, Cd, Zn, Cu, etc.).
E. minutumwas rare at C79 but was one of the dominant species at
C85, making this taxon metal sensitive. Similar findings were
registered by Rushforth et al. (1981), who found that Cymbella minuta
(synonym of E. minutum) showed preference for low concentrations
of Ag, Mn, Cd, Ni etc. Ivorra et al. (1999) also registered C. minuta at
the reference site but not at the polluted or the extremely polluted
site. E. minima showed higher relative abundances at C93. Ivorra et al.
(1999) also counted more Navicula seminulum (S. seminulum) at
polluted sites than at less polluted ones. Therefore, E. minima/N. se-
minulum is another metal tolerant species.
At C79 some taxa presented abnormal valves, mainly F. capucina
var. rumpens but also Fragilaria cf. crotonensis and A. minutissimum.
Scanning electron microscopic photos of F. capucina var. rumpens
reveal asymmetrical abnormal and bent valves with notched or
incised valves leading to alterations in striae patterns (Fig. 5).
F. capucina var. rumpens was counted at sites C79 (12%) and C85
(7%) but more than half of the valves counted showed structural
deformitiesat C79.The teratologiesat C79sitewere9 timesmorethan
at C85. Fragilaria cf. crotonensis was quite abundant at C79 and some
teratologies were also found, but was rare at C85 (Table 5).
Different species or infraspecific categories of F. capucina or other
Fragilaria species have been registered as dominant taxa in metal
polluted situations just as in the present study. Ivorra et al. (1999)
found, among other taxa, F. capucina as one of the dominant species,
while Gold et al. (2002) reported the increment of Fragilaria
crotonensis but also G. parvulum andPinnularia sp.within a transferred
community to a metal polluted site from a reference site (non
polluted). In a paleolimnology study Cattaneo et al. (2004) found that
F. capucina var. rumpens and Fragilaria cf. tenera dominated in
sediments corresponding to a Cu contamination period and to the
initial increase of other metals (Zn, Cd and Fe). These two taxa do not
usually dominate in the plankton or in the benthos as those authors
state and as also confirmed in this study. These taxa were rapidly
displaced by others (Achnanthes minutissima synonym of A. minutis-
simum) as metal pollution became a mixture of metals. Hirst et al.
(2002) concluded the opposite, this is, that F. capucina var. rumpens
was a sensitive taxon to some metals (Pb, Cu, Zn and Cd) being most
abundant at sites with low metal concentration. Our results and those
of other authors indicate that F. capucina var. rumpens is capable of
bearing high concentrations of several metals.
A. minutissimum (the most abundant taxon at C79 and C85) showed
only one abnormal valve at each site. The presence of abnormal valves,
especially of F. capucina var. rumpens in both polluted sites (C79 and
C85) was, most probably, due to metal contamination. Frustule
deformities and changes in cell size at metal contaminated sites have
been registered on several occasions. Gold et al. (2003) found twisted
frustules of F. capucina var. gracilis at two Cd and Zn polluted sites, but
with low relative abundances (b3%). Cattaneo et al. (2004) reported
a 3% relative abundance of valve deformities of Fragilaria cf. tenera,
F. capucina var. rumpens and Eunotia sp. These deformities were sig-
nificantly related to Cu concentration. Dickman (1998) found 2% of de-
formed benthic marine diatoms (F. capucina, Achnanthes hauckiana and
Diatoma vulgare) at a site with high concentrations of Cd, Cr, Cu, Ni, Pb
and Zn. McFarland et al. (1997) also found morphologically abnormal
Fragilaria valves and concluded that these deformities could be indi-
cators of high dissolved metal concentrations in streams.
It's interesting to note that several recent papers show distorted
valves of the genus Fragilaria in different geographical areas (Europe,
North America, and Asia) both freshwater and marine, and subjected to
different mixtures of metals, but none of them showed such high
relativeabundancesasinthepresentwork.Althoughthistaxonisableto
survive well (high relative abundances) at metal polluted sites, some
protective mechanism seems to fail resulting in the development of
distorted Fragilaria specimens. Several mechanisms have been devel-
oped by algae for tolerating metals at a cellular level: decrease in the
number of binding sites at the cell surface, uptake inhibition, and
development of exclusion or internal detoxifying methods like metal
accumulationinpolyphosphategranules(Twiss,1992).FisherandJones
(1981) in Gold et al. (2003) have suggested that metals inhibit normal
membrane function and reduce silicic acid uptake and amino-acid
synthesis leading to abnormalities during diatom silica wall formation.
Thislastmechanismisprobablyresponsibleforthedeformitiesfoundin
the present study.
Despite the hypothesis that high metal concentration is the major
cause for the diatom teratologies found, in future studies in this area,
other organic toxic substances (inducers of teratologies, too) such as
pesticides and herbicides should be analysed because agriculture is an
important local activity.
6. Conclusions
Tailing samples, stream sediments, surface waters and diatoms
were analysed downstream from an abandoned mine in Portugal. The
5633
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636

Page 15

studied area revealed interesting geochemical and biological aspects.
Total concentration of heavy metals in stream sediment samples
showed a decreasing metal gradient along the Coval da Mó (C232,
C79) and Fílvida (C85) streams located in the old mining area. This
stream presented a very high metal concentration when compared
with the samples collected in Vouga river and also in the background
area (Serra da Freita). The application of the consensus-based guide-
lines by using the Probable Effect Concentration values suggests that
CovaldaMóstreampresentsahighleveloftoxicity.Existenceofadown-
stream metal gradient concentration in stream sediments, and the lower
concentration of trace metals in the surface water samples suggest that
thedispersionismainlymechanical(presenceofpyrite,galenaandspha-
leriteinthestreamsamples)andthatthecarbonatephasesactasimpor-
tant factors to neutralize the media.
The sequential extraction method was able to show the relevance
of metal adsorption onto mineral surfaces, and the major importance
of the exchangeable phases and the residue on metal fixation. It must
beemphasizedthattheresultsofsomesecondaryphasessuchasangle-
site and smithsonite could be an important source of metals.
The geochemical results show a very toxic environment, especially
between C232 and C79. The diatom community revealed an increase
in species richness and diversity along the decreasing metal gradient
(from C232 to C85). The mixture of metals and their high concentra-
tions in stream sediments near the mine (C232) was too toxic to allow
a stable diatom community development. Further downstream (C79),
the decrease in metal concentrations to lower levels (two orders of
magnitude) permitted the growth of diatoms, many of which with de-
formations particularly in F. capucina var. rumpens (61% at C79 and 7%
at C85).
Considering that many environmental parameters can induce
changes in community structure as well as morphological abnorm-
alities, the interpretation of the biological results must have a physical
and chemical support in order to find the main causes for the biota's
responses. In the present study the environmental leading force is the
very high concentration of the metals. Therefore we suggest that dis-
torted valves of diatoms, in this case, indicate metal contamination.
Acknowledgements
The authors are grateful to the Aveiro University and the ELMAS
and GeoBioTec Research Center for their financial support during this
study. We thank the Geosciences Department for assistance in the
transport and chemical analyses and the Biology Department for
providing a laboratory for diatom study. We gratefully acknowledge
the anonymous reviewers whose comments and suggestions signifi-
cantly improved the manuscript.
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.scitotenv.2009.06.047.
References
Adshead-Simonsen P, Murray GED, Kushner DJ. Morphological changes in the diatom,
Tabellaria flocculosa, induced by very low concentration of cadmium. Bulletim of
Environmental Contamination and Toxicology 1981;26:745–8.
Almeida SFP. Utilização das diatomáceas na avaliação da qualidade das águas doces.
Aveiro: PhD Dissertation (Biology), Universidade de Aveiro; 1998. 524 pp.
ASTM. American Society for testing materials. Annual Book of ASTM Standards, Water
Environmental Technology, vol. 11.01.; 1984.
Austin A, Deniseger J. Periphyton community changes along a heavy metals gradient in
a long narrow lake. Environ Experim Botany 1985;25(1):41–52.
Baffico G, Wenzel M, Schimmele M, Pedrozo F, Diaz M, Koschorreck M, Neu T. Commu-
nity structure and photosynthetic activity of epilithon from a highlyacidic (pH ≤ 2)
mountain stream in Patagonia, Argentina. Extremophiles 2004;8:463–73.
Barber HG, Carter JR. Observations on some deformities found in British diatoms. Micro-
scopy 1981;34:214–26.
Besch WK, Ricard M, Cantin R. Benthic diatoms as indicators of mining pollution in
the Northwest Miramichi river system, New Brunswick, Canada. Hydrobiologia
1972;57:39–74.
Beutell A, Matzke M. Die Konstitution von Zinkblende, Wurtzit und Hauerit. Cbl. Min
1915:263–72.
Bird G, Brewer P, Macklin M, Balteanu D, Driga B, Serban M, Zaharia S. The solid state
partitioning of contaminant metals and As in river channel sediments of the mining
affected Tisa drainage basin, northwestern Romania and eastern Hungary. Appl
Geochem 2003;18:1853-1595.
Blanck H, Wangberg SA, Molander S. Pollution-induced community tolerance — a new
ecotoxicological tool. In: Cairns Jr J, Pratt JR, editors. Functional Testing of Aquatic
Biota for Estimating Hazards of Chemicals. ASTM StP 988. Philadelphia: American
Society for Testing and Materials; 1988. p. 219–30.
BorgF,HedlundM.GeochemicalstudyoftheoldminingareaintheDistrictofAveiro.MSc.
Lulea Department of Environmental Engineering, University of Technology; 2001.
62pp.
BouletMP,LarocqueACL.Acomparativemineralogicalandgeochemicalstudyofsulphide
mine tailingsat two sites in New Mexico, USA. Environ Geol 1998;33:130–42.
BradshawPD,ClewsRD,WalkerJL.Explorationgeochemistry.Rexdale,Ontario:Barringer
Research; 1972.
Breault RF, Reisig KR, Barlow LK, Weiskel PK. Distribution and potential for adverse
biological effects of inorganic elements and organic compounds in bottom sediment,
lower Charles River, MassachusettsU.S. Geological Survey Water-Resources Investi-
gations Report 00-4180; 2000. 70 pp.
Brigham ME. Elemental chemistry of streambed sediments of the St. Croix River Basin.
Water-Resources Investigations Report 02-4087U.S Department of Interior/U.S.
Geological Survey; 2002. 19 pp.
Butcher JT,Stewart PM, Simonc TP. A benthic community index for streamsin the north-
ern lakes and forests ecoregion. Ecological Indicators 2003;3:181–93.
Cabral JACN. Relatório Acerca das Minas do Braçal, Palhal, Carvalhal e Penna no Districto
Administrativo de Aveiro. Lisboa: Imprensa Nacional; 1859. 87 pp.
Cardoso Fonseca E, Ferreira da Silva E. Application of selective extraction in metal-
bearing phases identification: a South European case study. Journal of Geochemical
Exploration 1998;6:203–12.
Cardoso Fonseca E. Emploi de l'extraction chimique sélective séquentielle et détermina-
tiondesphases-supportduPbetduZnenmilieusilico-alumineuxlors del'alteration
supergène: exemple du prospect de Sanguinheiro (SE Aveiro — Portugal. Comuni-
cação dos Serviços Geológicos de Portugal 1982;68(2):267–83 in French.
Carter J. A new Eunotia and its great morphological variations under stress caused by a
habitat loaded with cooper salts. In: Ouvragae dédie à H. Germain. M. Ricard O.
Koelts, Koenigstein (Eds), 1990.
CattaneoA,AsioliA,ComoliP,MancaM.Organismsresponseinachronicallypollutedlake
supports hypothesized link between stress and size. Limnology and Oceanography
1998;43:1938–43.
Cattaneo A, Couillard Y, Wunsam S, Courcelles M. Diatom taxonomic and morphological
changesasindicatorsofmetalpollutionandrecoveryinLacDufault(Québec,Canada).
Journal of Paleolimnology 2004;32:163–75.
Chaussepied M. Manuel des Methods de Prevelement et d'Analises: Caracterisitiques
Physochimiques et Hydrobiologiques', Vol.1. Brest: Reseau National d'Observations
de la Qualité du Milieu Marine; 1977.
Cox EJ. What is the basis for using diatoms as monitors of river quality? In: Whitton
BA, Rott E, Friedrich G, editors. Use of algae for monitoring rivers. Austria: Univ.
Innsbruck; 1991.
CSST. Consensus-Based Sediment Quality Guidelines. Recommended for use and
application. Interium Guidance. Wiscosin: Dept. of Natural Resources WT-732;
2003. 35 pp.
Deniseger J, Austin A, Lucey WP. Periphyton communities in a pristine mountain stream
aboveandbelowheavymetalminingoperations.FreshwaterBiology1986;16:209–18.
DGRAH. Udometeorologia, Anuário dos Serviços Hidráulicos Direcção Geral
dos Recursos e Aproveitamentos Hidráulicos. Lisboa: Ministério das Obras Públicas;
1981.
Dickman MD. Benthic marine diatom deformities associated with contaminated
sediments in Hong Kong. Environment International 1998;24(7):749–59.
EdelgaardI,DahlströK.RiskassessmentforcontaminatedsitesinEurope.In:FergusonC,
Kasamas H, editors. Policy frameworks 2. Nottingham, Denmark: LQM Press; 1999.
p. 29–39.
Falasco E, Bona F, Badino G, Hoffmann L, Ector L. Diatom teratological forms and envi-
ronmental alterations: a review. Hydrobiologia 2009;623:1-35.
Ferreira AMPJ. Dados geoquímicos de base de sedimentos fluviais de amostragem de
baixa densidade de Portugal Continental: estudo de factores de variação regional.
Tese de Doutoramento da Universidade de Aveiro 2000; 226 pp.
Ferreira da Silva E, Matos J, Patinha C, Reis P, Cardoso Fonseca E. The effect of unconfined
mine tailings on the geochemistry of soils, sediments and surface waters of the
Lousal area (Iberian Pyrite Belt, Southern Portugal. Land Degradation and Devel-
opment 2005:213–28.
Ferreira Da Silva E, Patinha C, Reis P, Cardoso Fonseca E, Matos JX, Barrosinho J, Santos
Oliveira JM. Interaction of acid mine drainage with waters and sediments at the
Corona stream, Lousal mine (Iberian Pyrite Belt, Southern Portugal. Environmental
Geology 2006;50:1001–13.
Fisher NS, Jones GJ. Effects of copper and zinc on growth, morphology, and metabolism
of Asterionella japonica (Cleve). J Exp Mar Biol Ecol 1981;51:37–56.
Foster PL. Metal resistances of Chlorophyta from rivers polluted by heavy metals.
Freshwater Biology 1982;12:41–61.
Gensemer RW. Role of aluminium and growth rate on changes in cell size and silica
content of silica-limited populations of Asterionella ralfsii var. Americana (Bacillar-
iophyceae). Journal of Phycology 1990;26:250–8.
5634
E. Ferreira da Silva et al. / Science of the Total Environment 407 (2009) 5620–5636