Soil pollution by oxidation of tailings from toxic spill of a pyrite mine
M. Simón, , F. Martín, I. Ortiz, I. García,
J. Fernández, E. Fernández, C. Dorronsoro and J. Aguilar
Dpto. de Edafología y Química Agrícola, Facultad de Ciencias,
Universidad de Granada, 18071 Granada, Spain
1. Introduction
2. Methods
3. Results and discussion
3.1. Tailings
3.1.1. Physical characteristics and chemical composition
3.1.2. Oxidation and solubilization of the pollutants
3.2. Soil pollution
3.2.1. From 0 to 10 cm in depth
3.2.2. From 10 to 30 cm in depth
4. Conclusions
References
1. Introduction
Pyrite mines have been worked for centuries in the province of Seville (southern
Spain), especially in the Aznalcóllar mining district, on the eastern
edge of the Iberian Pyrite Belt (Carvalho, 1976). The mineral phases consisted
of different sulfides such as pyrite (83.1%), sphalerite (5.4%), galene (2.1%),
chalcopyrite (1.4%) and arsenopyrite (0.9%), as well as minor amounts of bournonite,
boulangerite, nuffieldite, jaskolskiite and numerous trace metals ( Almodovar
et al., 1998). On 25 April 1998, the walls of two contiguous ponds containing
the ore-processing residues from the pyrite mine located in Aznalcóllar
(Spain) broke open, and approximately 36Å~105 m3 of polluted water (solution
phase) and 9Å~105 m3 of toxic tailings (solid phase) spilled into the
Agrio and Guadiamar River Basin (Fig. 1).
The toxic spill spread some 50 km downriver (the solid phase spread 37 km, and
the solution phase the entire 50 km), affecting some 40 km2 of farmland.
.
On 4 May (9 days after the spill), we studied seven sectors in the affected area, analysing tailings, polluted water, and contaminated as well as uncontaminated soils (Simón et al., 1999). The principal pollutants of the soils were: Zn, Pb, Cu, As, Sb, Bi, Cd and Tl. The range of total contamination of each element was extremely broad, as penetration of the tailings depended on the soil characteristics. Most of the Cu, Zn and Cd penetrated the soil in the solution phase of the spill, while the other elements penetrated mostly as part of the solid phase. Nevertheless, in these types of tailings, as a result of drying and, consequently, aeration, sulfides oxidize to sulfates, the pH falls markedly due to the formation of sulfuric acid, and the pollutants solubilize ( Förstner and Wittmann, 1983). Therefore, with future rains, this acidic solution would infiltrate the soil and aggravate the soil-pollution problem.
In the present study, we continued to monitor the concentrations of the pollutants
in the tailings and soils in five of the seven aforementioned sectors, to ascertain
the oxidation rate of the tailings as well as the increase in soil contamination
over time. Because the tailings were removed from the surface of the soils,
this monitoring ended on 22 July 1998 (88 days after the spill).
Five sectors were studied along the basins of the Agrio and Guadiamar Rivers:
near the mine (MIN), at the point of the spill; Soberbina (SOB), at 5.5 km from
the spill; Aznalcázar (AZN), at 21 km; Quema (QUE), at 29 km; and Pescante
(PES), at 36 km (Fig. 1). In each sector,
a square plot was laid out (25Å~25 m). At each corner and in the centre
of the plot, samples were taken of tailings as well as of soil at 0–10
cm and at 10–30 cm in depth. All samples, categorized according to origin
(tailings and the two soil depths), were air dried and screened (2 mm mesh size).
Next, 250 g of each sample category from the five sampling points per plot were
mixed and homogenized for laboratory analysis. Following this procedure, given
that the five sectors had been sampled on 4 May 1998 ( Simón et al.,
1999), we continued sampling each plot on three different dates: 20 May (25
days after spill), 4 June (40 days after spill) and 22 July 1998 (88 days after
spill). The soils of the plots were classified into four categories according
to the World Reference Base For Soil Resource ( FAO, ISRIC, ISSS, 1998): Calcaric
Fluvisols (MIN and QUE), Eutric Arenic Fluvisol (SOB), Calcaric Regosol (AZN)
and Calcaric Hiposalic Regosol (PES).
Particle-size distribution was determined by the pipette method after the elimination
of organic matter with H2O2 and dispersion with sodium hexametaphosphate (Loveland
and Whalley, 1991). Bulk-density data were obtained using the clod method. The
pH was measured potentiometrically in a 1:2.5 soil–water suspension. The
total sulfur was analysed by dry combustion with a LECO instrument. A saturated
extract of the tailings was prepared, and the sulfates were precipitated as
BaSO4, following the guidelines of the US Salinity Laboratory Staff (1954).
Samples of the tailings and soils, very finely ground (<0.05 mm), were digested
in strong acids (HNO3+HF+HCl). In each digested sample and saturated extract
of the tailings, Cu, Zn, Cd, As, Pb, Sb, Bi and Tl content were measured by
ICP-MS with a PE SCIEX ELAN-5000A spectrometer. The accuracy of the method was
corroborated by analyses (six replicates) of a standard reference material:
SRM 2711 (soil with moderately elevated trace element concentrations; Gills
and Kane, 1993) ( Table 1). For the
statistical analysis, the StatView 4.02 program was used.
The climate of this area is typically Mediterranean (hot, dry summers; cold, wet winters; temperate autumns and springs with variable rainfall). The rainfall data from 4 May to 22 July 1998 (late spring–early summer) in the five sectors studied are shown in Table 3. The data for MIN and SOB were taken from the Aznalcóllar weather satation (very close to the point of the spill), while the data for AZN, QUE and PES were taken from the Aznalcázar weather station (at 21 km from the spill).
3.1.1. Physical characteristics and chemical composition
The data from the 4th May 1998, the sampling date on which the tailings were
still relatively moist and had not yet undergone major oxidation, were taken
as the reference to estimate the original characteristics of the tailings. According
to these data, the particle sizes of the tailings (Table
2) proved to be related to distance from the spill (Table 4), the fine-silt
augmenting (P<0.1) and the coarse-silt (P<0.1) as well as the sand (P<0.05)
fractions diminishing. These relationships did not present a higher statistical
significance level, because beyond AZN (at 21 km from the spill), the particle
size tended to remain relatively constant. The clay fraction was not related
to distance from the spill, whereas the fine silt content of the tailings proved
to be negatively related to the bulk density (P<0.05), and positively to
the sulfur content (P<0.01).
The pollutant contents differed in behaviour according to the element (Table 3). Thus, the content in As, Bi, Tl and Pb showed a positive relationship (P<0.01) with the fine-silt and S contents of the tailings (Table 4). Also, Sb was positively related to these parameters, although with lower significance (P<0.1). The Cu content tended to increase with higher fine-silt and S content, but without statistical significance. Meanwhile, Zn and Cd, with minor, uneven variation over the distance of the spill, were not related to any of the parameters of the tailings nor with any of the other contaminants. This lack of relationship could be explained by the fact that part of the Zn, Cd and Cu in the tailings precipitated (presumably as hydroxides) from the solution phase of the spill, in which their concentration was relatively high (Simón et al., 1999).
3.1.2. Oxidation and solubilization of the pollutants
In these types of tailings, as a result of drying and, consequently, aeration,
sulfides oxidize to sulfates (Nordstrom, 1982), the pH falls markedly due to
the formation of sulfuric acid ( Stumm and Morgan, 1981) and the formerly insoluble
pollutants partly solubilize ( Rogowski and Caruccio). Because the sulfate ions
that form remain soluble in the acidic solution, the oxidation rate can be estimated
from the ratio between sulfur as soluble sulfate (Ss) in the saturation extracts
and total sulfur (St) of the tailings. Given that on the 4th May 1998, the tailings
remained relatively moist and their oxidation rate (Table
3) was more strongly related to moisture than to other characteristics (
Simón et al., 1999), the differences in the oxidation rate of the tailings
was estimated from the relationship Ss/St for the sampling on 20 May 1998, when
the tailings had reached relatively uniform dryness. This oxidation rate clearly
increased with the distance from spill (Table
3), maintaining a positive relationship with the fine-silt and total-sulfur
contents in the tailings (P<0.01), and negative one with the sand, coarse-silt
and bulk density (P<0.05; Table 4).
In all the sectors, this oxidation process was very rapid, reaching the highest
concentration of sulfates and lowest pH values on 20th May (25 days after spill;
Fig. 2). After the 20th May, the sulfate
content stopped rising, indicating a slowing of the oxidation rate, and in fact
this content fell between the 20th May and 4 June, implying a greater loss of
sulfate ions by leaching than those formed by oxidation during this period.
This pattern of concentration in sulfate ions was paralleled by that of water-soluble
Zn, Cu and Cd (Fig. 2), with the solubilization
on the 20th May reaching between roughly 20% (sector MIN) and 65% (sectors AZN,
QUE and PES) of the total Zn and Cd present in the tailings, and between approximately
8% (sector MIN) and 45% (sectors AZN, QUE and PES) of the total Cu. This reflects
a swift and intense oxidation of sulfides bonded to these elements (such as
sphalerite or chalcopyrite). In addition, the marked fall in the pH values must
also have prompted the solubilization of the hydroxides of these elements precipitated
from the solution phase of the spill.
Nevertheless, other sulfurs present in the tailings, such as arsenopyrite,
galene or jaskolskiite oxidated more slowly and less intensely, as reflected
by the fact that the greatest concentration values of soluble As, Pb and Bi
in the tailings were reached on 4 June (40 days after spill; Fig.
3) and that these values represent only a minor fraction with respect to
the total (the highest values being approx. 2.5% for As, 2.0% for Bi, 0.4% for
Sb and 0.1% for Pb). Soluble Tl showed a singular distribution. On 4 May, values
were highest, but clearly differed between sectors according to the drying and
oxidation rates of the tailings (Simón et al., 1999). On 20 May, values
dropped abruptly in all the tailings and remained very low until the end of
the monitoring period. This trend appears to indicate that the solubilized Tl
rapidly leached by rains did not accumulate in the tailings.
The total concentrations of S, Zn, Cu, Cd and Tl in the tailings of all the
sectors progressively diminished over time (Table
3). The less soluble elements such as As, Pb, Sb and Bi, although increasing
in concentration on 20 May (apparently owing to a greater leaching of the more
soluble elements), also decreased afterwards; i.e. the elements solubilized
during the oxidation and acidification of the tailings leached with the rainwater,
accumulating in the underlying soil and thereby exacerbating the soil pollution.
3.2.1. From 0 to 10 cm in depth
Given that between 25 April (date of the spill) and 4 May (date of the first
sampling) no rain fell (Table 3), the
concentration in the different pollutants of the soils sampled on 4 May would
be attributable only to the soil concentrations prior to the spill (background)
and the amount that penetrated the soil from the solution and solid phases of
the toxic spill (initial pollution). These concentrations were highly variable
( Table 5) because the range of the
initial pollution was extremely broad, as the penetration of the solid phase
depended on the soil characteristics; values being especially high in PES, where
the amount of tailings that penetrated the soil reached 150 g kg-1 (Simón
et al., 1999). On the contrary, the increase of these elements in the soils
sampling on 20 May or later, can be ascribed to the oxidation of the tailings,
solubilization of the pollutants in the rainwater and infiltration into the
soils (oxidative pollution). Thus, the rate of oxidative pollution of each element
(OPi) on a given date can be calculated by the difference in the concentrations
between the first soil sampling (on 4th May) and any subsequent one.
The oxidative pollution of the soils between 0 and 10 cm in depth augmented with time (Fig. 4), and by 22nd July 1998, had more than doubled the overall concentration of the pollutants in the first 10 cm of the soils, except in PES where the initial pollution was very heavy (Table 5). The majority of this contamination in Zn, Cd, Cu, Sb and Tl (65–90%) occurred between the 4th and 20th May ( Fig. 4), the period in which the strongest oxidation and solubilization of the sulfides bonded to these elements in the tailings took place. During this period, the rainfall ranged from 45 mm (in the sectors closest to the spill) to 63 mm (in the sectors farthest from the spill). Meanwhile, most of this contamination in As, Pb and Bi (70–90%) occurred between the 20th May and 4th June ( Fig. 4), when the oxidation and solubilization of the sulfides bonded to these elements in the tailings were highest and the total rainfall from the date of the spill reached between 60 and 89 mm (in the sectors closest and farthest from the spill, respectively). Between the 4th June and 22nd July, the oxidative pollution increased very little in the absence of rainfall.
As a result of this intense oxidative pollution, on the 22nd July 1998 (88 days after spill) the first 10 cm of the soils of MIN, SOB, AZN and QUE reached or exceeded the maximum concentration permitted for agricultural soils of As (50 mg kg-1; NMHPPE, 1991, in Alloway, 1995) and Tl (1 mg kg-1; Kabata Pendias and Pendias, 1992). The soils of AZN and QUE exceeded the maximum permitted concentrations in Zn and Cu (720 and 190 mg kg-1, respectively; NMHPPE, 1991, in Alloway and Van). In the soils nearest the spill (MIN and SOB), where the initial pollution and subsequent oxidation and solubilization rates of the tailings were lower, the increases in the Zn and Cu content were weaker and the above-mentioned toxicity limits were not reached. Meanwhile, the soil concentration in Cd, Pb, Bi and Sb was considerably raised by the oxidative pollution, but in no sector, except PES, did Cd and Pb exceed the maximum level allowed for agricultural soils (12 and 530 mg kg-1, respectively; NMHPPE, 1991, in Alloway and Van). In PES, this maximum was surpassed by the initial pollution in all elements, except in Cd. For Bi and Sb, no reference levels are available.
3.2.2. From 10 to 30 cm in depth
The oxidative pollution on the 22nd July 1998 from 10 to 30 cm in depth was
considerably less than that of the first 10 cm of the soils (Table
6), averaging 95% less in Tl, 80–85% less in Cd, Zn, Cu and Pb, and
70–80% less in As. In all sectors, the concentrations in the different
pollutants were far lower than the maximum permitted for agricultural soils.
Consequently, the pollutants tended to remain in the first 10 cm of the soils
without seriously contaminating either the subsoil or the groundwater.
The characteristics of the tailings deposited on the soils proved to be related
to distance from the spill, the fine-silt fraction and the sulfur, As, Pb, Bi,
Tl and Sb content increasing considerably and the bulk-density decreasing. The
Zn, Cd and Cu content were not related to distance from the spill. The oxidation
rate of the tailings, acidification and solubilization of the pollutants clearly
increased with the distance from the spill. In each sector, the water-soluble
sulfates, Zn, Cd and Cu increased very rapidly and intensely. Meanwhile, the
increases in water-soluble As, Bi, Sb, Pb and Tl were far lower and less rapid
in the case of As, Bi and Pb.
The pollutants solubilized during the oxidation and acidification of the tailings
leached with the rainwater, intensifying the soil pollution with time. As a
result of this oxidative pollution, 25 days after the spill, the first 10 cm
of the soils in the middle and lower sectors of the basin exceeded the maximum
concentration permitted for agricultural soils of Zn, Cu and Tl; meanwhile,
40 days after spill, all the soils reached or exceeded the maximum permitted
concentrations in As and Tl. Consequently, a rapid removal of the tailings is
essential and should begin in the middle an lower sector of the basin. The oxidative
pollution from 10 to 30 cm in depth was considerably less than that of the first
10 cm of the soils and the concentrations in the different pollutants were far
lower than the maximum permitted for agricultural soils. Therefore, after the
removal of the tailings, the first 25–30 cm of the soils could be homogenized
by ploughing as a means of reducing the overall concentration of pollutants.
Alloway, 1995B.J. Alloway, Editor, Heavy metals in soils, Blackie Academic &
Professional, London (1995), p. 368.
Almodovar et al., 1998G.R. Almodovar, R. Saez, J.M. Pons and A. Maestre , Geology
and genesis of the Aznalcóllar massive sulphide deposits, Iberian Pyrite
Belt, Spain. Miner Deposita 33 (1998), pp. 111–136. Abstract-GEOBASE
Caruccio and Geidel, 1978F.T. Caruccio and G. Geidel , Geochemical factors affecting
coal mine drainage quality. In: F.W. Schaller and P. Sutton, Editors, Reclamation
of drastically disturbed lands, ASA-CSSA-SSSA, Madison, Wisconsin (1978), pp.
128–148.
Carvalho, 1976D. Carvalho , Consideraçoes sobre o vulcanismo da regiao
de Cercal-Odemira. Suas relaçoes com a Faixa Piritosa. Comun Serv Geol
Port 60 (1976), pp. 215–238.
FAO, ISRIC, ISSS, 1998FAO, ISRIC, ISSSIn: World Reference Base For Soil Resources.
Roma (1998), p. 91.
Förstner and Wittmann, 1983U. Förstner and G.T.W. Wittmann In: Metal
pollution in the aquatic environment, Springer-Verlag, Berlin (1983), p. 486.
Gills and Kane, 1993T.E. Gills and J.S. Kane Certificate of analysis. Standard
reference material 2711, National Institute of Standards and Technology, Gaithersburg,
USA (1993).
Kabata Pendias and Pendias, 1992A. Kabata Pendias and H. Pendias Trace elements
in soils and plants, CRC Press, Boca Raton, Florida (1992).
Loveland and Whalley, 1991P.J. Loveland and W.R. Whalley , Particle size analysis.
In: K.A. Smith and Ch.E. Mullis, Editors, Soil analysis: physical methods, Marcel
Dekker, New York (1991), pp. 271–328.
Nordstrom, 1982D.K. Nordstrom , Aqueous pyrite oxidation and consequent formation
of secondary iron minerals. In: J.A. Kitrick, D.S. Fanning and L.R. Hossner,
Editors, Acid sulfate weathering, Soil Sc Soc Am, Madison, WI (1982), pp. 37–56.
Rogowski et al., 1977A.S. Rogowski, H.B. Pionke and J.G. Broyan , Modeling the
impact of strip mining and reclamation processes on quality and quantity of
water in mined areas: a rewiew. J Environ Qual 6 (1977), pp. 237–244.
Abstract-EMBASE | Abstract-Compendex
Simón et al., 1999M. Simón, I. Ortiz, I. García et al.,
Pollution of soils by the toxic spill of a pyrite mine (Aznalcóllar,
Spain). Sci Total Environ 242 (1999), pp. 105–115. SummaryPlus | Full
Text + Links | PDF (749 K)
Stumm and Morgan, 1981W.Y. Stumm and J.J. Morgan In: Aquatic chemistry: an introduction
emphasizing chemical equilibria in natural waters, John Wiley & Sons, New
York (1981), p. 218.
US Salinity Laboratory Staff, 1954US Salinity Laboratory StaffIn: Diagnosis
and improvement of saline and alkaline soils, Handbook 60, US Department of
Agriculture, Washington DC (1954), p. 160.
Van Den Berg et al., 1993R. Van Den Berg, C.A. Dennenman and J.M. Roels , Risk
assessment of contaminated soil: proposal for adjusted, toxicologically based
Dutch soil clean-up criteria. In: F. Arendt, G.J. Annokkée, R. Bosman
and W.J. van der Brink, Editors, Contaminated Soils '93, Kluwer Academic Publisher,
London (1993), pp. 349–364.