Document Sample
dam Powered By Docstoc
					                                                              Science of the Total Environment 407 (2009) 5761–5771

                                                                  Contents lists available at ScienceDirect

                                                      Science of the Total Environment
                                                j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / s c i t o t e n v

Combination of sequential chemical extraction and modelling of dam-break wave
propagation to aid assessment of risk related to the possible collapse of a roasted
sulphide tailings dam
Rafael Pérez-López a,b,⁎, Reinaldo Sáez b, Antonio M. Álvarez-Valero b,c, José Miguel Nieto c, Gaetano Pace d
  Institute of Environmental Assessment and Water Research, IDÆA-CSIC, Jordi Girona 18, 08034, Barcelona, Spain
  Department of Geology, University of Huelva, Campus ‘El Carmen’, 21071, Huelva, Spain
  Andalusian Institute of Earth Sciences (IACT), CSIC-University of Granada, Campus 'Fuentenueva', Faculty of Sciences, 18002, Granada, Spain
  Advanced Computer Systems SpA, Environment Division, Via della Bufalotta 378, 00139, Rome, Italy

a r t i c l e         i n f o                             a b s t r a c t

Article history:                                          The Sotiel–Coronada abandoned mining district (Iberian Pyrite Belt) produced complex massive sulphide ores
Received 13 March 2009                                    which were processed by flotation to obtain Cu, Zn and Pb concentrates. The crude pyrite refuses were roasted for
Received in revised form 15 July 2009                     sulphuric acid production in a plant located close to the flotation site, and waste stored in a tailing dam. The
Accepted 16 July 2009
                                                          present study was focused on the measurements of flow properties, chemical characterization and mineralogical
Available online 15 August 2009
                                                          determination of the roasted pyrite refuses with the aim of assessing the potential environmental impact in case
                                                          of dam collapse. Chemical studies include the determination of the total contaminant content and information
Almagrera tailing dam                                     about their bio-availability or mobility using sequential extraction techniques. In the hypothetical case of the
Risk analysis                                             tailing dam breaking up and waste spilling (ca. 4.54Mt), a high density mud flow would flood the Odiel river
Waste spillage                                            valley and reach both Estuary of Huelva (Biosphere Reserve by UNESCO, 1983) and Atlantic Ocean in matter of a
Numerical modelling                                       couple of days, as it was predicted by numerical simulations of dam-break waves propagation through the river
Bio-available contaminants                                valley based on quasi-2D Saint-Venant equations. The total amount of mobile pollutants that would be released
                                                          into the surrounding environment is approximately of 7.1·104 t of S, 1.6·104 t of Fe, 1.4·104 t of As, 1.2·104 t of Zn,
                                                          1.0·104 t of Pb, 7.4·103 t of Mn, 2.2·103 t of Cu, 1.5·102 t of Co, 36t of Cd and 17 t of Ni. Around 90–100% of S, Zn, Co
                                                          and Ni, 60–70% of Mn and Cd, 30–40% of Fe and Cu, and 5% of As and Pb of the mobile fraction would be easily in
                                                          the most labile fraction (water-soluble pollutants), and therefore, the most dangerous and bio-available for the
                                                          environment. This gives an idea of the extreme potential risk of roasted pyrite ashes to the environment, until now
                                                          little-described in the scientific literature.
                                                                                                                                        © 2009 Elsevier B.V. All rights reserved.

1. Introduction                                                                                  inates large areas of land or water bodies, causing extensive harm to the
                                                                                                 affected ecosystems.
    Mining is one of the main industrial activities occurring in almost every                        Social alarm related to the risk of the existing tailing dams has
part of the world (ICOLD, 1996). Through ore extraction and processing,                          grown since recent catastrophes occurred. In April 1998, one of the
large amounts of waste material are produced and stored, much of it being                        most remarkable cases of dam breakage happened in Aznalcóllar
of a highly toxic nature. Of waste storage sites, mine tailing dams are more                     mining district (Iberian Pyrite Belt (IPB), southern Spain) when the
vulnerable than other retention structures and constitute the greatest                           impoundment wall of a pond holding mine tailings from the Los
threat because of their: (1) unstable nature (principally of earth                               Frailes copper–lead–zinc–silver mine collapsed, releasing 7 million
construction), (2) location (often situated near sensitive aquatic ecosys-                       m3 of toxic waste in the form of heavy metal-rich sludge and acidic
tems), (3) large number (in the tens of thousands world-wide), and (4)                           water (López-Pamo et al., 1999; Simón et al., 1999). Almost two years
generally poor to non-existent maintenance (especially after closure of                          later, in January 2000, a similar catastrophe occurred when approxi-
mining activities) (ICOLD, 2001). Little is known of their potential harmful                     mately 100,000 m3 of tailings, comprising of cyanide and heavy
impact on the environment, except in a few well-publicised cases, when                           metal-rich sediments from a gold-recovery operation in northern
catastrophic waste spillage from breached or over-topped dams contam-                            Romania near the town of Baia Mare, overflowed into the Vaser, Lapus,
                                                                                                 and Somesu rivers, and subsequently into the Tisza and Danube rivers
                                                                                                 (Korte et al., 2000; Soldán et al., 2000).
 ⁎ Corresponding author. Institute of Environmental Assessment and Water Research,
CSIC, Jordi Girona 18, 08034, Barcelona, Spain. Tel.: +34 95 921 9826; fax: +34 95 921
                                                                                                     Over the past 30 years other tailing dams incidents occurred in
9810.                                                                                            these and other countries or states of the world, e.g., Sweden, Italy,
    E-mail address: rafael.perez@dgeo.uhu.es (R. Pérez-López).                                   Portugal, United Kingdom, Papua New Guinea, Kyrgyzstan, Australia,

0048-9697/$ – see front matter © 2009 Elsevier B.V. All rights reserved.
5762                                       R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771

Ghana, French Guyana, Bolivia, Colorado, Arizona, although they did                 In order to estimate the pollutant speciation in studied roasted pyrite
not receive much publicity. A review of tailings dam failures could be              refuses, we applied the simplest and most standardized sequential
found in Rico et al. (2008a,b). In most cases, collapsed tailing dams               extraction method proposed by the European Community Bureau of
derived from either lead–zinc–copper–silver or uranium or gold                      Reference (BCR) (Ure et al., 1993), and improved in later studies
mining operations. These unpleasant events evidence that the pre-                   (Rauret et al., 1999; Sahuquillo et al., 1999).
vention actions should receive much more attention and funds than                       Given the huge volume of production of roasted pyrite ashes and
post-disaster remediation activities.                                               the limited research on their potential impact on the environment, the
    The e-EcoRisk project (www.e-ecorisk.info), funded by the                       information presented in this paper may be used to design possible
European Commission, 5th Framework Program, includes the analysis                   remediation strategies at roasted pyrite ashes storage systems
of risk associated with hazardous spillage from large scale tailing                 distributed around the world (Salomons, 1995; Gupta et al., 1996).
storage systems. In the scope of this EU project, a sulphuric acid
production plant located in Almagrera industrial complex was                        2. Materials and methods
selected as one of the Spanish test sites to be fully characterized for
risk analysis for three reasons: (1) residues produced and stored in                2.1. Site description
dams in this type of complex display a little-known chemical nature,
with little information on the threat posed by them to the                              Sotiel–Coronada mine is one of the most important polymetallic
environment; (2) dam stability requires a continuous monitoring                     sulphide mining districts at the IPB. Ores are fine grained showing
and control during emplacement, construction and operation, but the                 pervasive intergrowth of pyrite, chalcopyrite, sphalerite and galena as
owner company of this industrial complex declared bankruptcy in                     the main components. Arsenopyrite, tetrahedrite–tennantite, mene-
2001 and the facilities were suddenly abandoned; and (3) in case of                 ghinite, bournonite, pyrrhotite, marcasite, cubanite, enargite, stannite,
dam failures, the zone that may be affected potentially has a huge                  bismuthinite, kobellite, gudmundite, boulangerite, cassiterite, mag-
ecological value (RAMSAR and NATURA 2000 conservation areas).                       netite and gold have also been identified as accessory minerals (García
    Sulphuric acid is the most produced chemical reagent in the world.              de Miguel, 1990). Main gangue minerals include quartz, chlorite and
Worldwide production in 2001 was around 165 million tons. Pyrite                    carbonate (ankerite–siderite). Original reserves of more than 120 Mt
(FeS2) is often used as an ore for production of sulphuric acid via                 have been reported (Santos et al., 1996) including both the Sotiel and
roasting and SO2 recovery with water. In theory, this process is                    Migollas orebodies. Mining began during the Roman Empire. The
environmentally clean and produces a waste known as roasted pyrite                  modern mining operation started during the second half of the 19th
ashes that can be used as iron ore in steel, brick, paint and cement                century, later being closed in 1939 by The United Alkali Co. The mine
industries since they are composed of high concentrations of                        was opened again in October 1983 by Almagrera S.A., and then closed
hematite. However, this sub-product is not inert but contains con-                  in 2001.
siderable amounts of pollutants elements (e.g., Fe, Cu, S, Zn, Pb and As)               The Almagrera project involved the construction of a huge
that can be very mobile in environmental conditions and reduce                      industrial complex located near the Calañas village in the Huelva
drastically its applicability. Despite this, and with the exception of one          Province, SW of Spain (Fig. 1). This complex includes both flotation
paper (Lin and Qvarfort, 1996), we have not found other works in the                and sulphuric acid production plants. In the flotation plant, poly-
present literature that justify the risk of this type of residues to the            metallic sulphide ores from Sotiel–Coronada mine were processed to
environment.                                                                        obtain Cu–Zn–Pb concentrates. The coarse pyrite-rich fraction from
    In Almagrera industrial complex, there is a tailing pond containing             flotation refuses was processed to produce sulphuric acid, oleum and
roasted pyrite ashes whose management has not been controlled or                    copper sulphate, by roasting and SO2 recovery. This process was
monitored by any public entity since 2001. There are suspicions that                enhanced by catalytic improvement (Morales, 1986). Two tailing
the tailings dam does not meet present-day stability requirements.                  ponds were also built, one for pyritic waste from flotation process
Within e-EcoRisk project, Rico et al. (2008a,b) and Salgueiro et al.                (sulphide tailing) and the other for roasted pyrite ashes from
(2008) proposed simple empirical equations to describe the potential                sulphuric acid production (oxide tailing) (Fig. 1).
collapse risk and outflow hydraulic behaviour of tailing ponds based                     The project was designed for 730,000 metric tons per year,
on the available information on historic failures. These authors                    producing 90,000 metric tons of base-metal concentrates. The
analysed the case of Almagrera roasted pyrite ashes-tailing dam and                 sulphuric acid plant has capacity for processing 221,000 metric tons
deduced that its potential risk of collapse is similar to that of the               of pyrite giving 300,000 t of sulphuric acid. During the plant operative
historical case of Aznalcóllar dam. The possible breakage and waste                 period, 3 cubic hectometres of roasted pyrite refuses were produced
spilling would be a new catastrophe for both surrounding environ-                   (Morales, 1986). The sulphide tailing dam is similar in origin and
ment and inhabitants of this region because the Almagrera tailing dam               characteristics to those that collapsed in the Aznalcóllar mining
is also located at the IPB.                                                         district, and therefore, the potential harmful impact on the environ-
    The main aims of this study are: (1) to assess the polluting capacity           ment that would cause the waste spillage is already widely known
of Almagrera roasted pyrite ashes through physical, chemical and                    (Álvarez-Valero et al., 2009). However, few investigations focused on
mineralogical studies; (2) to develop a dynamic quasi two-dimen-                    wastes from the pyrite roasting have been reported in the literature,
sional model to simulate and predict the outflow (discharge), in case                and the only existing study found describes the case of a sulphuric acid
of tailing dam collapse, and the pathways and spatio-temporal                       plant located in Sweden (Lin and Qvarfort, 1996).
dispersion of waste into the environment; and (3) to combine both
BCR-sequential extraction data and numerical simulations in order to                2.2. Sample collection and preparation
establish the potential environmental impact at regional scale after a
hypothetical waste dam collapse. Chemical characterization includes                    A total of 50 samples of roasted pyrite ashes (approx. 1 kg each)
not only the determination of total content of pollutants but also an               were collected in the oxide tailing dam using a polypropylene shovel,
estimation of the chemical binding to wastes, which determines the                  and subsequently transferred to clean polypropylene bags in August
mobility and bio-availability of the elements (Weisz et al., 2000; Yu               2004. The waste dumping in the pond was carried out from a filling
et al., 2001). The sequential extraction procedure is the most common               point. The processing system and ore did not record any change during
technique used to calculate the different forms of metals in e.g.                   the activities of the industrial complex (Morales, 1986). Hence, the
sediments, soils, wastes whereby several reagents are utilized                      main waste features were constant in time and the only factor that
consecutively to extract operationally defined phases in a sequence.                 conditioned a differentiation in the deposit may have been the
                                                                                                                                                                                                                                      R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771
Fig. 1. Location map of the Almagrera industrial complex (Huelva, SW Spain). Open circles are the location of sampling points in the oxide tailing pond and solid grey circles are those points subjected to sequential extraction.

5764                                                 R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771

granulometry. Like in fan deposits in sedimentary basins, the grain-                               performed in those four representative samples of the proximal, lateral,
size variation in a longitudinal section (finer fraction toward the distal                          middle and distal part, with respect to the filling point (Fig. 1).
zone) is similar to those recognized in vertical profile with depth                                     Samples from the sequential extraction were analysed by Activa-
(finer fraction toward the top part). Thus, in order to sample all                                  tion Laboratories Ltd (ACTLABS, Ontario, Canada), accredited under
possible variations, the sampling network was designed in several                                  ISO 9001 and 9002. A total of 36 elements (Ba, Al, K, Mg, Mn, Si, Ag, As,
profiles (parallel and perpendicular) to the main filling channel                                    Be, Bi, Ca, Cd, Ce, Co, Cr, Fe, Cu, Li, Mo, Na, Ni, P, Pb, Sb, S, Se, Sn, Sr, Te,
(Fig. 1). For it, only one depth was necessary (ca. 40 cm). Nevertheless,                          Ti, Tl, U, V, W, Y and Zn) were determined by atomic emission
one previous study of the sediment structure of the dam carried out                                spectroscopy with inductively coupled plasma (ICP-AES). To validate
through both boreholes and electric tomographies showed an evident                                 the results, the analysis sequence consisted of calibration standards,
homogeneity in all waste (Andalucia Region Government, unpub-                                      standard solutions analysed as an unknown (quality control solu-
lished data). Sample points were located by Global Position System                                 tions), method blanks, replicate solutions and two certified reference
(GPS) on a digital orthophoto with a scale of 1:10.000. In the                                     solutions: riverine water (SLRS-4) from the National Research Council
laboratory, the samples were stored under frozen conditions (− 4 °C).                              of Canada (NRCC) and natural water (NIST 1640) from the National
All analyses were carried out within two weeks after sampling.                                     Institute of Standards and Technology (Gaithersburg, USA).
                                                                                                       The most polluting and mobile metals are discussed: S, Fe, As, Cd,
2.3. Analytical methods                                                                            Co, Cu, Mn, Ni, Pb and Zn. Hereafter, the term metal includes metals
                                                                                                   and metalloids (i.e. As). Although the sequential extraction procedure
    As an input for risk analysis, roasted pyrite ashes from Almagrera                             is usually applied to assess metal mobility, S was also discussed within
oxide dam have been characterized in chemical and mineral composi-                                 this group since it is the most abundant element in leachates from IPB
tion. Analyses were carried out in the Central Research Services of the                            mining wastes. In order to verify the accuracy of the procedures, two
University of Huelva. The chemical composition was conducted on the                                types of checks on the results were performed in the laboratory: (1)
Bruker S4 Pioneer X-ray fluorescence spectrum (XRF) using Rh target                                 an external check of Cd, Cr, Cu, Ni, Pb and Zn through the triplicate
and 4 kW maximum power. The mineralogical characterization was                                     analysis of a standard sediment reference material (BCR-701, the
carried out by X-ray diffraction (XRD, powder method) using a Bruker                               European Community Bureau of Reference); and (2) an internal check
diffractometer (model D8 Advanced). Working conditions were slit                                   of As, Cd, Co, Cu, Mn, Ni, Pb and Zn by comparing the sum of the 4 steps
fixed at 12 mm, Cu Kα monochromatic radiation, 20 mA and 40 kV.                                     (acid-soluble + reducible + oxidizable + residual) from the sequential
Samples were run at a speed of 0.3° 2θ/min (5–60°). Fluorite as internal                           extraction procedure with the metal analysis from XRF. This last
reference material was used to determine semi-quantitative miner-                                  internal check consisted of calculating the percentage recovery of the
alogical composition. The high predominance of main minerals                                       sequential extraction procedure as follows (Eq. (1)):
obscures some other accessory minerals, which are important for
environmental assessment. In order to complete the mineralogical                                                     Step 1 + Step 2 + Step 3 + Step 4
                                                                                                   Recoveryð%Þ =                                       × 100                     ð1Þ
characterization, samples were also observed by means of scanning                                                            XRF total element
electron microscopy equipped with an energy dispersive system
(SEM-EDS) for the chemical microanalysis (JEOL JSM-5410).
                                                                                                   3. Flow numerical simulation
2.4. Sequential extraction method and quality control
                                                                                                       The systematic analysis of flow behaviour and flood effects on river
    The modified BCR-sequential extraction procedure is summarised in                               valley from roasted pyrite waste in case of dam failure was
Table 1 and full details can be found elsewhere (Rauret et al., 1999;                              numerically modelled using the 1D DAMBRK computer code (Fread,
Sahuquillo et al., 1999). The environmental hazard of the roasted pyrite                           1985). The mathematical model hereafter presented computes the
ashes depends directly on their capacity to release easily potentially                             resultant outflow hydrograph (discharge as a function of time) and
toxic metals into water. Fraction 1 consists of exchangeable metals and                            simulates the dam break flood wave formation and subsequent
those soluble in water or in slightly acidic conditions. It is the fraction                        downstream progression. The modelling was developed for water
with the most labile bond to the wastes, and therefore, the most                                   containing high sediment concentration, such as those observed in
dangerous and bio-available for the environment. Fractions 2 and 3 can                             debris and mud flows resulting from landslides mainly in mountain
also indicate a threat depending on environmental conditions. Fraction 2                           regions. These rapid movements of saturated soil, rock and other
represents metals bound to Fe and Mn oxides that can be released if                                debris may cause harmful disasters, due to their particular behaviour
conditions change from oxic to anoxic state. Fraction 3 is made up with                            such as the impact force or the ability to carry large boulders.
metals bound to sulphides and organic matter (the latter is not included                               The governing equations of the model in which the routing process
in these residues), which may be released under oxidizing conditions.                              is based are the complete one-dimensional Saint-Venant equations
Finally, fraction 4 corresponds to those metals strongly associated with                           that are coupled with internal boundary equations representing the
crystalline structures of minerals, which are therefore unlikely to be                             flow through structures such as dams as well as external boundary
released from the wastes. Since the sediment structure shows the                                   equations for the upstream and downstream end of the routing reach.
homogeneous infill of the oxide tailing, the sequential extraction was                              The original equations developed by Barre de Saint-Venant (1871)

Table 1
The modified BCR-sequential extraction procedure used for speciation of metals.

Step            Solid phases                                       Extractant (1 g of dry solid)                                             Shaking time and temperature
F1              Water/acid soluble and exchangeable                40 mL of 0.11 M CH3COOH                                                   16 h at room temperature
F2              Reducible                                          40 mL of 0.5 M HONH2·HCl (pH 2)                                           16 h at room temperature
F3              Oxidizable                                         10 mL of 8.8 M H2O2 (pH 2) then 10 mL of 8.8 M H2O2                       1 h at room temperature and 1 h at 85 °C
                                                                   (pH 2) cool, add 50 mL of 1 M NH4OAc (pH 2)                               1 h at 85 °C
                                                                                                                                             16 h at room temperature
R               Residuala                                          10 mL of aqua regia (12 M HCl + 15.8 M HNO3 in the ratio 3:1)             Heating on hot plate to dryness

F1 (easily soluble fraction), F2 (reducible fraction), F3 (oxidizable fraction) and R (residual fraction).
    Digestion of the residual material is not a step of the BCR protocol.
                                                 R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771                                 5765

lack the lateral dimension; hence, in order to provide numerical                          and full details can be found elsewhere (see e.g. DAMBRK User Manual,
simulations compatible with the observed bank erosion measure-                            1990).
ments, it is necessary to reformulate the conceptual model in order to                        The method applied for the determination of basin geometrical
include the lateral sediment input and section enlargement. It is in this                 parameters was the 3D restitution from topographic maps previous to
sense that the model is quasi-2D; the closure equations and the main                      dam construction. These maps were originally made by Almagrera
assumptions, most notably the direction of the flow, are those of the                      topographers at scale 1:10000. Digitalised contour lines were marked
one-dimensional conceptual model. The quasi-2D model suitable to                          for shape and volume restitution, as well as, for the integration of
tackle this bank erosion problem is characterized by the following                        basin parameters in a Geographical Information System (GIS; ArcView
conservation of mass equation (Eq. (2a)) and conservation of mo-                          3.2© and CAD software). Data obtained by means of this approach are
mentum equation (Eq. (2b)):                                                               consistent with those reported by Almagrera. The flow parameters of
                                                                                          roasted pyrite ashes were determined at the University of Huelva.
∂A   ∂Q                                                                                   Instability of a material upon slope arises when the shear strength or
   +    −q = 0                                                                   ð2aÞ
∂t   ∂x                                                                                   maximum resistance to shear stress is exceeded by a downslope
       2                                                                                stress. This situation depends on internal cohesion, angle of internal
     ∂ Q               
∂Q      A        ∂h                                                                       friction and liquid limit of the material. Cohesion is produced by the
   +       + gA     + S0 = 0                                                     ð2bÞ
∂t     ∂x        ∂x                                                                       interlocking of particles that enables the material to rest at an angle
                                                                                          and internal friction is the resistance of particles to slide across each
where t is time coordinate, x is measured along the slope of the river in                 other. Both parameters are obtained from triaxial tests. Liquid limit
the downstream direction, A is the cross sectional area of flow (L2), Q is                 represents the minimum water content at which soil particles flow
the discharge (L3T− 1), q is the lateral inflow or outflow per lineal                       under their own weight. This value is determined by the three-point
distance along the channel (inflow is positive and outflow is negative in                   Casagrande method, the 3 mm rod formation and shrinkage mould
sign; L2T− 1), g is gravitational acceleration (LT− 2), h is the water surface            techniques (McBride, 1993).
elevation and S0 is friction slope due to boundary resistance. The                            The simulation was defined for various scenarios, corresponding to
expanded Saint-Venant equation constitutes a system of partial                            various rupture conditions (25 and 100% downstream discharge), and
differential equations with two independent variables, x and t, and                       for each scenario, it was possible to predict the following: time of
two dependent variables, h and Q. The solutions of these equations are                    arrival of the flood wave at a certain point in space, flow rate and
too complex to be solved by analytical methods, and numerical methods                     amount of sediment at a certain point in space and a certain time,
are absolutely necessary. The numerical method used to solve the model                    height of the flow rate in space and time and total area covered by the
is based on the well-known MacCormack predictor–corrector numerical                       sediment. The geometric information of the river bed in each point in
scheme, a second-order explicit finite difference two-step scheme based                    space (slope and cross-section area) was obtained using DEM (Digital
on two partial time steps for each Δt (McCormack, 1971; Ferreira and                      Elevation Map) in GIS software.
Leal, 1998). On the basis of the model developed, the following input
data are required for the functionality of the software (Table 2):                        4. Results
(1) geometric information of dam and waste volume, (2) dam's material
properties (grain size, average water content, dry density, solid particle                4.1. Quality control of analysis
weight, liquid limit, porosity, cohesion and internal friction angle) and
(3) geometric information of the river bed. The influence of input data                        The analytical performance of the laboratory and analytical
on parameters of Eq. (2) for debris or mud flows is summarised below                       variability were evaluated by both applying the procedures to BCR-
                                                                                          701 and calculating the recovery percentage between sequential
Table 2
                                                                                          extraction and total element analysis, as described in the section on
Main input parameters for the setup of the discharge model.
                                                                                          experimentation. On the one hand, the results obtained for extractable
Basin information            Length of dam                        762 m                   amounts together with indicative or certified values following the
                             Width of dam                         109 m
                                                                                          procedures for the standard reference material are shown in Table 3.
                             Height of dam                        35 m
                             Surface of basin                     289,227 m2              The application of a paired t-test (α = 0.05) showed that certified (or
                             Waste volume                         2,667,771 m3            indicative) and obtained values were not significantly different. This
Fluid information            Height level of the fluid in          32 m                    indicated that the quality of analytical extraction data was ensured.
                             the basin                                                    Also, comparison of the sum of the four steps average values from the
                             Height level of the fluid in the      0.0005 m
                                                                                          sequential extraction with the metal content by XRF through Eq. (1)
                             Average water content of the         37.6%                   revealed good percentage recoveries for As, Cd, Co, Cu, Mn, Ni, Pb and
                             sediment                                                     Zn, indicating the reliability of the data. The average recovery value for
Sediment information         Grain size D80                       20–65 µm                all metals was 97 ± 29% (Table 3).
                             Dry density                          1.7 g/cm3
                             Solid particle weight                3.98 g/cm3
                             Liquid limit                         27%                     4.2. Chemical and mineralogical characteristics
                             Porosity                             57.2%
                             Cohesion (kPa/cm2)                   8.4–10.6                   The average values from chemical and mineralogical characteriza-
                             Internal friction angle              23°–28°                 tion of the roasted pyrite samples are reported in Table 4. Roasted
River bed information        Total distance                       80 km
                             Slope in fluvial section              from 2.5 to 0.1%
                                                                                          pyrite ashes are mainly composed of hematite and poorly crystalline
                             (55 km)                                                      Fe oxides, occurring typically as pellets (Fig. 2a). Quartz and
                             Width in fluvial section              b100 m                  occasionally Fe sulphate (mainly minerals of jarosite group), barite
                             Depth in fluvial section              b2 m                    and pyrite (as the remains of non-oxidized sulphides) may appear as
                             Slope in estuarine section           from 0.1 to 0.001%
                                                                                          accessory minerals. Iron oxides pellets locally show a porous or
                             (25 km)
                             Width in estuarine section           N200 m                  spongy texture including minor fine grained minerals (Fig. 2b). These
                             Depth in estuarine section           6–10 m                  samples are characterized by the high iron content, which comprises
    Brief descriptive summary of the river geometric data. Total information used to
                                                                                          around 80 wt.% of the waste. The presence of S content around 7 wt.%
compute the model along 80 km has been obtained using a Digital Elevation Map taken       explains that the iron oxides cannot be reused in the metallurgical
from Andalucia Region Government (2005).                                                  industry. Inert elements include Si, Al, Ti, Ca, Mn and Ba. Heavy metals
5766                                                 R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771

Table 3
Quality control of data using (1) measured, certified and indicative values for extractable amounts in certified reference material BCR-701 and (2) comparative results (% recovery) of
average analyses of the BCR sequential extraction and XRF total analysis.

Control check                                  Elements
                                               As               Cd               Co               Cr               Cu               Mn                Ni               Pb               Zn
BCR-701 (mg/kg)
F1            Obtained value                   –                7.05 ± 0.09      –                2.10 ± 0.15      48.0 ± 1.5       –                 14.1 ± 0.1       2.55 ± 0.01      184 ± 7
              Certified value                   –                7.34 ± 0.35      –                2.26 ± 0.16      49.3 ± 1.7       –                 15.4 ± 0.9       3.18 ± 0.21      205 ± 6
F2            Obtained value                   –                3.21 ± 0.02      –                47.0 ± 0.9       138 ± 3          –                 27.5 ± 1.1       121 ± 3          98 ± 6
              Certified value                   –                3.77 ± 0.28      –                45.7 ± 2.0       124 ± 3          –                 26.6 ± 1.3       126 ± 3          114 ± 5
F3            Obtained value                   –                0.09 ± 0.07      –                129 ± 13         46.9 ± 4.9       –                 13.8 ± 3.7       9.5 ± 1.8        43.3 ± 13.3
              Certified value                   –                0.27 ± 0.06      –                143 ± 7          55.2 ± 4.0       –                 15.3 ± 0.9       9.3 ± 2.0        45.7 ± 4.0
R             Obtained value                   –                0.05 ± 0.01      –                52.3 ± 7.7       32.6 ± 3.8       –                 31.3 ± 0.8       12.0 ± 0.7       69.6 ± 0.1
              Indicative value                 –                0.125 ± 0.075    –                62.5 ± 7.4       38.5 ± 11.2      –                 41.4 ± 4.0       11.0 ± 5.2       94.6 ± 12.2
Recovery (%)                                   115              73               87               –                98               134               66               68               136

and metalloids such as Zn, Pb, Cu, S, As and Sb are present in quantities                         in proximal and lateral zones, unlike As, which presents an opposite
up to 6 wt.%. These high polluting elements are bound to Fe sulphates                             behaviour. With respect to the total concentration for each metal, the
and oxides.                                                                                       average percentage in the mobile fraction follows the order: S (81%) N
                                                                                                  Mn (43%) N As (38%) N Cd (31%) N Pb (28%) N Co (23%) N Cu (18%) N Zn
4.3. Sequential extraction data                                                                   (17%) N Ni = Fe (3%) (Fig. 3). The most potentially toxic fraction for the
                                                                                                  environment, i.e. the bio-available fraction (water soluble or F1), is
   The environmental hazard of roasted pyrite ashes not only                                      composed of S (75% of total), Mn (29%), Co (20%), Cd (19%), Zn (15%),
depends on the total metal content but also on their mobility and                                 Cu (7%), Ni (3%) and Fe = Pb = As (1%). The total mobile content of Ni
bio-availability determined by the sequential extraction. The                                     is leached in the bio-available fraction. This waste shows high As (27%
extracted percent values of S, Fe, As, Cd, Co, Cu, Mn, Ni, Pb and Zn in                           of total), Pb (22%), Mn (9%), Cd (7%) and Cu (5%) contents related to
the four samples of roasted pyrite ashes in each sequential extraction                            the reducible fraction (bound to oxides, i.e. F2), as well as minor
step are shown in Table 5 and the average values represented in Fig. 3.                           contents of metals in the oxidizable fraction (bound to sulphides, i.e.
The total concentrations of elements extracted in the mobile phases                               F3): As (11% of total) N Pb (6%) N Cu = Cd = Mn (5%) N S (4%).
(based on the sums of the first three fractions, i.e. F1 + F2 + F3) are in
the following abundance order in mg/kg:                                                           4.4. Spatio-temporal propagation of dam-break waves
 - S (12042) N As (3521) N Fe (3481) N Zn (1363) N Mn (1296) N Cu
                                                                                                      The sudden collapse of the roasted pyrite dam would trigger a high
   (512) N Pb (116) N Co (18) N Cd (5) N Ni (2), for the sample collected
                                                                                                  density mud flow, with similar effects to those of the Aznalcóllar dam
   in the proximal part.
                                                                                                  break disaster. It would be expected to spread mainly through Odiel
 - S (14174) N As (4916) N Fe (2857) N Mn (1157) N Zn (523) N Cu
                                                                                                  river, after being reached by Batán stream of approx. 1 km in length
   (416) N Pb (196) N Co (8) N Cd (5) N Ni (2), for the sample collected
                                                                                                  (Fig. 1). The Odiel river Basin is the largest drainage basin in Huelva
   in the lateral part.
                                                                                                  province, with an area of about 2300 km2. The Odiel river starts in the
 - S (20789) N Zn (4923) N Pb (4914) N Fe (4704) N Mn (1972) N As
                                                                                                  Sierra de Aracena and, together with the Tinto river, flows into a
   (824) N Cu (402) N Co (56) N Cd (13) N Ni (6), for the sample
                                                                                                  coastal wetland known as the Ría of Huelva estuary, which contains
   collected in the middle part.
                                                                                                  one of the most important marsh ecosystems in Europe. The Odiel
 - S (16029) N Pb (3805) N Zn (3735) N Fe (3479) N As (2900) N Mn
                                                                                                  river drops over 600 m on its way from the headwaters to the mouth,
   (2068) N Cu (618) N Co (48) N Cd (10) N Ni (5), for the sample
                                                                                                  about 140 km away. The tailing dam is located in the central part of the
   collected in the distal part.
                                                                                                  basin and the pathway followed by the discharge would be of approx.
    Although total metal concentration is relatively uniform in the                               80 km in length divided into two subsections: (1) a clearly fluvial
tailing, some heterogeneities exist with respect to metal mobility. The                           section of 55 km and (2) an estuarine section of 25 km until the exit to
main singularity is that Zn and Pb contained in the mobile phases                                 the Atlantic Ocean. In the estuarine section, the salt marshes
present much higher concentrations in middle and distal zones than                                associated to the right margin of the Odiel river were declared as a

Table 4
Chemical and mineralogical characteristics of the roasted pyrite ashes.

Major elements (wt.%)
           Fe2O3             SO3                SiO2                 CaO              Al2O3            TiO2             MgO               MnO              K2O              Na2O
Mean       80.71             6.98               3.31                 1.21             1.19             0.58             0.42              0.36             0.14             0.12
SD         5.38              3.98               1.11                 0.94             0.61             0.27             0.21              0.13             0.07             0.04
Range      65.62–89.64       0.33–20.73         1.20–5.82            0.15–4.60        0.32–2.85        0.02–0.93        0.15–1.06         0.12–0.80        0.03–0.28        0.05–0.26

Trace elements (mg kg− 1)
           Pb                Zn                 As                   Cu               Sb               Sn               Ni                Co               V                Cr               Bi
Mean       11696             11484              6894                 2830             535              504              271               162              79               70               68
SD         3959              4284               2012                 822              125              210              123               46               34               20               38
Range      7789–24,170       5635–25,443        4175–13,240          1659–5823        0–788            172–931          87–726            76–260           3–132            18–112           28–228

Mineralogical composition (wt.%)
           Hematite          Quartz                                  Fe oxides                         Fe sulphates                       Sulphides                         Barite
           62                21                                      12                                2                                  2                                 1

A summary statistic – mean, standard deviation and range – is provided for waste composition from the analysis of 50 samples.
                                                       R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771                                              5767

                                                                                                Fig. 3. Average percentage of S, Fe, As, Cd, Co, Cu, Mn, Ni, Pb and Zn extracted in each
                                                                                                step of the sequential extraction procedure for samples of roasted pyrite ashes.

                                                                                                Odiel river marshes as a humid zone to be protected on an interna-
                                                                                                tional scale.
                                                                                                    Based on the simulation, the outflow hydrograph of the residue if
                                                                                                the dam collapse occurs is shown in Fig. 4. The computations indicated
                                                                                                that the total duration of significant outflow or emptying the reservoir
                                                                                                is about 2 h, a similar time in both rupture scenarios considered (25
                                                                                                and 100% discharge). The maximum peak value of discharge is
                                                                                                900 m3/s to 40 min and 200 m3/s to 1 h for the scenarios of 100 and
                                                                                                25% discharge, respectively. In both cases, the discharged residue is
                                                                                                propagated as flood waves through the downstream river valley
                                                                                                (Fig. 5). The case of 100% discharge has much higher impact, both in
                                                                                                the velocity with which it reaches the downstream area and,
                                                                                                especially in the amount of sediment. As observed in Fig. 5, for the
                                                                                                same time coordinate, dam-break flood waves in the case of 100%
                                                                                                discharge transport higher concentration of sediment and at a slightly
                                                                                                faster rate than those of 25% discharge.
                                                                                                    Regardless of the discharge percentage, the amplitude of dam-
Fig. 2. Back-scattering electron composition images of roasted pyrite waste. (a) General        break waves is attenuated gradually as it propagates downstream until
texture and composition of the waste. (b) Internal structure and composition of a               reaching the estuarine area. In the estuary, the rapid increase in width
hematite pellet.
                                                                                                and depth along with the decrease of the slope in the channel sharply
                                                                                                reduce the amplitude and dissipate the wave propagation model.
Biosphere Reserve by UNESCO in 1983, famous for its high                                        From the dynamic point of view, Estuary of Huelva is strongly
ornithological diversity and for the presence of halophytic vegetation.                         influenced by tides, a feature that can be known from its funnel-
The Spanish Government signed the RAMSAR agreement to include                                   shaped mouth. In general, this morphology determines that the
                                                                                                propagation of tidal wave in the interior of the estuary depends on
Table 5                                                                                         two opposing factors, on one hand, the convergence effect causing the
Results obtained for BCR sequential extraction analysis for samples of roasted pyrite ashes
                                                                                                acceleration of tidal currents upstream, and on the other hand, the
from the proximal (PP), lateral (LP), middle (MP) and distal (DP) part of the oxide tailing.

Step Sample     Elements (mg/kg)
                S       Fe        As      Cd     Co      Cu     Mn     Ni    Pb     Zn
F1   PP      12,042     850   40           2      16      202  591   2    0   1160
     LP       14,174    329   77           1       6      107  280   2    0    357
     MP      18,565    2112   54          10      51      269 1884   6  112  4332
     DP      13,873     772   49           6      42      227 1665   5  116  3296
     Average 14,663    1016   55           5      29      201 1105   4   57  2286
F2   PP            0  2485 2368            1       1      146  482   0   66     95
     LP            0  2398 3264            2       1      101  552   0   96     63
     MP         774    1891 654            2       3       75   54   0 3709    219
     DP         750   2336 2317            2       4      248  262   0 3131    179
     Average    381   2278 2151            2       2      142  337   0 1750    139
F3   PP            0    146 1113           1       1      164  223   0   50    108
     LP            0    129 1575           1       1      208  326   0  101    103
     MP        1450     701 116            1       2       58   34   0 1093    372
     DP        1406     370 534            1       2      143  141   0  558    261
     Average     714    337 835            1       2      143  181   0  450    211
R    PP        3649 126,984 4999          13      63     2076 2288 107 6094  6495
     LP        6512 133,308 7661          16      93     2823 3869 98 8964 12,105
     MP        1934 134,210 3028          24     128     1628  901 130 3520 21,389
     DP        2625 130,484 3862          16     146     2580 1620 136 4237 11811
                                                                                                Fig. 4. Outflow hydrograph from hypothetical collapse of Almagrera roasted pyrite
     Average   3680 131,246 4888          18     108     2277 2169 118 5704 12,950
                                                                                                tailing dam: 100% discharge (solid line) and 25% discharge (dash line).
5768                                             R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771

                                                                                          Younger et al., 2002). Moreover, mining wastes are highly porous and
                                                                                          unsaturated flow systems where there is a continuous atmospheric
                                                                                          oxygen renewal enhancing the sulphide oxidation and the subsequent
                                                                                          production of large amounts of AMD. The contaminants bound to
                                                                                          sulphides that are easily released in oxidizing conditions are those
                                                                                          extracted from the oxidizable fraction of the sequential extraction
                                                                                          (fraction 3). The presence of sulphides in most of IPB mining wastes
                                                                                          guarantees the existence of metals associated to this fraction that are
                                                                                          released with AMD.
                                                                                              Roasted pyrite ashes are, however, an atypical waste with respect to
                                                                                          remaining IPB mining wastes. The pyrite content is less than 2 wt.% since
                                                                                          this mineral was roasted for the sulphuric acid production. This explains
                                                                                          the low proportions of sulphur and metals associated with the
                                                                                          oxidizable fraction of the sequential extraction (approx. 5%) (Fig. 3).
                                                                                          Nevertheless, roasted pyrite refuses are also “acid mine drainage”
                                                                                          producing residues through a process not described previously in the
                                                                                          literature and that does not imply natural sulphide oxidation. The
                                                                                          abundance of S-bearing minerals in this waste (sulphides, Fe sulphates
                                                                                          and barite) is not enough to explain its total S percentage (ca. 7 wt.% de
                                                                                          SO3). This is because iron oxides pellets present a porous texture (Fig. 2)
                                                                                          with high surface area able to adsorb SO2 during the roasting process
                                                                                          (Lin and Qvarfort,1996). The adsorbed S represents 50% of the total S (ca.
                                                                                          3.5 wt.% de SO3) and is released readily when iron oxides react with
                                                                                          infiltrating water, which explains the high S concentrations leached in
                                                                                          the bio-available fraction of the sequential extraction (Fig. 3). The result
                                                                                          is the production of an acidic solution containing high sulphate
                                                                                          concentrations able to both dissolve and transport bio-available metals
                                                                                          in solution.
                                                                                              Acid mine waters are usually supersaturated in secondary minerals
                                                                                          that precipitate sequestering relatively high concentrations of poten-
                                                                                          tially toxic elements from the solution (Bigham et al., 1994; Webster
                                                                                          et al., 1998). Jarosite has been identified in the roasted pyrite ashes as
                                                                                          secondary mineral according to SEM-EDS observations. Likewise,
                                                                                          other hydrated metallic sulphate salts as secondary minerals (e.g.
                                                                                          epsomite, hexahydrite, copiapite, halotrichite, rozenite and coquim-
                                                                                          bite) are usually observed by SEM-EDS, albeit they are not easily
Fig. 5. Spatio-temporal propagation of dam-break waves through the river valley based
                                                                                          identifiable owing to the both low cristallinity and the many metals
on numerical simulations for the cases of (a) 100% discharge and (b) 25% discharge.       incorporated into the crystallographic network. As the iron oxides
                                                                                          make up about 75 wt.% of the roasted pyrite ashes, they are possible
                                                                                          candidates to produce leachates with high sulphate concentration
dissipation effect due to friction. In the Estuary of Huelva, convergence                 (fraction 1), which is favourable for the formation of the secondary
exceeds dissipation by friction, and hence the propagation model of                       salts. Most of these salts are temporary phases for metal retention
tidal wave is hypersynchronic, i.e. the tidal amplitude increases                         since they are highly soluble and become the secondary sources for
upstream before diminishing towards the river and tidal currents                          acidity, sulphate and metals through leaching with water. Therefore,
reach their maximum intensity in the central part of the estuary                          these minerals are also dissolved in the fraction 1 of the sequential
(Borrego et al., 1993). This strong tidal influence would control the                      extraction procedure. Some of them may also precipitate from
mixing processes between fluvial and marine water, as well as the                          solutions resulting during sulphide oxidation (fraction 3).
homogeneous dispersion of waste from the breakage of the roasted                              The tailing pond of roasted pyrite is exposed to weathering
pyrite dam into the estuary.                                                              conditions. The climate of the IPB is of a Mediterranean type, i.e. rainy
   The hypersynchronic tidal model of the Estuary of Huelva, in                           winters and warm–dry summers. This means that the most potentially
addition to being described in the literature, is also reflected in the                    dangerous fractions in this waste are the bio-available and, to a lesser
simulation carried out in the current study. For a time when dam-                         extent, oxidizable ones. As in all mining districts during warm periods,
break waves still move by river section, it is clearly observed in the                    the metallic sulphate salts originate from strong evaporation of the
estuarine section (last 25 km) an increase of the discharge values with                   metal-rich solutions (Olías et al., 2004). Moreover, the high porosity of
a maximum in the central zone, especially noticeable for the case of                      these wastes favours the continuous atmospheric oxygen input and the
25% discharge (Fig. 5b).                                                                  release of metals associated to sulphide (oxidizable fraction), which also
                                                                                          precipitate as sulphate salts. Whereas in rainy periods, metals bound to
5. Discussion                                                                             the bio-available fraction, made of both sulphate salts and S adsorbed
                                                                                          into iron oxides, are released into solution. The dissolution of soluble
   The main environmental problem deriving from IPB mining wastes is                      phases in rainwater is especially rapid with the first autumn rainfalls
a direct consequence of the high content in sulphide minerals, mostly                     (Cánovas et al., 2007). This process is cyclically repeated every year.
pyrite. Sulphides react readily with oxygen and water, and their                              In theory, the tailing pond is a hydrologically closed system where
weathering implies a series of chained geochemical and microbiological                    both acidity and metals are periodically recycled as a result of the
reactions whose result is the production of extremely acid drainages                      seasonal climatic variations. However, the dam is not totally water-
containing high concentrations of sulphate, metals and metalloids,                        tight and some metal-rich effluents coming from the leaching of this
known as Acid Mine Drainage (AMD) (Parker and Robertson, 1999;                            waste may reach the Odiel river (Sánchez-España et al., 2005).
                                                   R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771                                                5769

Nonetheless, the Odiel river is already contaminated by effluents from                          10–50 cm thick layer of waste would completely cover the total distance
other IPB mining districts when receiving the pollution load from                              of 80 km in approx. 1 day and 18 hours. After disaster, water-soluble
Almagrera industrial complex. As a summary, IPB hosts more than 80                             metals would be the first to be released: 6.6·104 t of S, 1.0·104 t of Zn,
massive sulphide deposits exploited (Sáez et al., 1999) and around                             5.0·103 t of Mn, 4.6·103 t of Fe, 9.1·102 t of Cu, 2.6·102 t of Pb, 2.5·102 t of
2 × 108 m3 of waste material. The acid leachates, rich in metals, derive                       As, 1.3·102 t of Co, 22t of Cd and 17 t of Ni (exact obtained values are
from the superficial oxidation of sulphides contained in these wastes                           shown in Table 6). This easily released amount of Pb, As, Mn, Zn, Cd, Co, S,
and are drained by the fluvial courses of the IPB, i.e. Tinto and Odiel                         Fe, Cu and Ni is respectively 9.6, 6.9, 3.1, 3, 2, 1.8, 1.1, 0.6, 0.5 and 0.4 times
rivers, causing their total pollution. Unexpectedly both rivers trans-                         higher than those transported by both rivers to the estuary during the
port the contaminant load to the Estuary of Huelva, affecting even the                         entire year, which further contribute to the pollutant load threatening
UNESCO-RAMSAR wetlands sites. Olías et al. (2006) calculated the                               the habitat of flora and fauna in protected coastal wetlands. In addition,
fluvial metal contribution to the Estuary of Huelva between 1995 and                            waste spilling would increase its exposure to weathering processes,
2003, showing that mean contaminants discharge is 7900, 5800,                                  enhancing the sulphide oxidation process and the release of metals
3500, 1700 and 1600 t/year of Fe, Al, Zn, Cu and Mn, respectively. Part                        associated with these minerals (oxidizable fraction, Table 6). Likewise, in
of contamination load is dispersed by coastal currents of the Gulf of                          the Estuary of Huelva the existence of reducing environments favours
Cádiz, with some metals reaching the Mediterranean Sea through the                             the precipitation of diagenetic pyrite (Monterde, 2004), which conse-
Strait of Gibraltar. This pollution has caused the extinction of the                           quently should also favour the release of metals associated to oxides
aquatic life of these rivers, with the exception of some acidophilic                           from the roasted pyrite particulate material (reducible fraction, Table 6).
microorganisms, and is producing serious biologic imbalances in fish                            To conclude, both the mobile metal content and flow properties of this
and bivalve molluscs from the estuarine environment (Funes et al.,                             waste provide evidence of its extreme contamination potential, which
2006). Moreover, Doñana National Park, Europe's largest nature                                 would have even greater impact if one considers some physical conse-
reserve and another internationally designated RAMSAR wetland site,                            quences to ecosystems, for example from the destruction of aquatic life
is situated next to Estuary of Huelva. This wildlife refuge was seriously                      by simple smothering of river or estuarine beds with sediment.
damaged during the Aznalcóllar disaster.                                                           This kind of information should be considered in the risk manage-
    The combination of metal mobility data from BCR-sequential                                 ment process conducted by local authorities. In fact, as a consequence of
extraction and numerical simulations of dam-break flood enables the                             this study, the roasted pyrite ashes-tailing dam was totally sealed to
prediction of a few hydrodynamic and chemical characteristics that                             avoid this possible environmental disaster by the Regional Environment
can be used for damage evaluation. Sequential extraction investiga-                            Authorities of the Andalucia Region Government during the second half
tions normally estimate the risk potential of waste based on relative                          of 2006. Given that roasted pyrite ashes are widely produced, we
comparisons. For this reason, the results (see Section 4.3) are                                strongly recommend the design of remediation strategies in the storage
described in terms of the percentages leached in each fraction with                            systems of this type of residues.
respect to the total quantity. Nevertheless, we intend to go further in
the application of extraction techniques by extrapolating the                                  6. On the relevance of the current approach
theoretical risk potential obtained in the laboratory to the real risk
potential on a regional scale, as reported by Pérez-López et al. (2008).                           The main goal of the e-Ecorisk project was to create a regional
In order to have an idea of the polluting capacity of the pyrite roasted                       enterprise network information management and decision-support
ashes for the surrounding environment, it is important to consider:                            system that could provide environmental and civil protection agencies,
the absolute average amount of mobile contaminants obtained in lab-                            and other relevant governmental and non-governmental organisations,
experiments (Table 5) and the total mass of waste released in case of                          with improved information and insight into the potential and actual
catastrophe. Taking both factors into account, the total amount of                             risks (impact) to the environment of large-scale industrial spills for
mobile contaminants that would be released in different environ-                               better decision-making. Although the current paper represents only one
mental conditions (total exposure of residues to water, oxygen or                              case study, results are being incorporated into a database including
reducing conditions) could be easily estimated.                                                societal, economic, legal, and political aspects relevant to environmental
    As the failure of the dam could happen in many ways and it is not                          risk of numerous study areas endangered by tailings dam spills. Within
possible to know the amount of flood discharged in the downstream,                              the e-EcoRisk project, case study sites selected to assess the risk related
various types of scenarios could be implemented, and thus, an                                  to the possible rupture or overflow of tailings dams are located in metal
emergency plan may be prepared according to each situation. In the                             mining regions of Portugal (Panasqueira and Aljustrel mines), Spain
worst case, total failure of the dam would release a total mass of waste of                    (Almagrera industrial complex and Riotinto mine), Italy (Gavorrano,
4.54 Mt, a value calculated from both volume estimated by GIS and                              Masua, Montevecchio, Monteponi and Campo Pisano mines) and Greece
density. Biological effects in the receiving environments would however                        (Stratoni and Olympias mines). In addition to the Almagrera roasted
be dependent on contaminant concentrations in waters that evolve                               pyrite ashes-tailing dam, the integration of geochemical analysis with
through the release of metals from solid phases. Table 6 shows the                             flow modelling has been implemented to the Montevecchio, Olympia
maximum amount of mobile contaminants in solution that would be                                and Panasqueira test sites. The Mediterranean region was chosen for the
released from pyrite roasted tailing pond considering a 100% discharge.                        sites because of the large number and size of active and abandoned mine
Under this scenario and plotting the temporal variation of height of the                       tailings dams to be found there, their often poorly maintained condition
dam-break wave as a function of distance (Fig. 6), one can observe as a                        and proneness to spillage, the paucity of information on the threat posed

Table 6
Total amount (tons) of bio-available, bound to oxides and bound to sulphides metals in roasted pyrite ashes tailing pond and that would be released into the environment in the worst
case of total dam failure (100% discharge).

Roasted sulphide (4.54 Mt)                   Elements (t)
                                             S              Fe             As             Cd           Co           Cu             Mn            Ni          Pb               Zn
Bio-available fraction                       66,572           4612            249         22           130           914           5016          17             259           10,380
Reducible fraction                             1729         10,340          9764           9            10           647           1532           0           7947               631
Oxidizable fraction                           3242            1528          3789           6             8           651            821           0           2044               957
Total impact                                 71,543         16,481         13,803         36           148          2212           7369          17          10,250           11,968
5770                                             R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771

          Fig. 6. Height profiles of dam-break waves through the river valley for the overtopping simulations in the worst case of total dam failure (100% discharge).

by them to the environment (particularly to nearby situated RAMSAR                        1.4·104 t of As, 1.2·104 t of Zn, 1.0·104 t of Pb, 7.4·103 t of Mn, 2.2·103 t of
and NATURA 2000 sites), and the highly vulnerable and sensitive nature                    Cu, 1.5·102 t of Co, 36t of Cd and 17 t of Ni. The most potentially toxic
of their environments to a mine tailings dam spill. The sites also rep-                   fraction for the environment, i.e. the bio-available fraction (water-
resent a diverse set of conditions, ranging from types of tailings ponds                  soluble pollutants), contains around 90–100% of S, Zn, Co and Ni, 60–70%
toxic metals (e.g., various base metals, uranium), to both inland and                     of Mn and Cd, 30–40% of Fe and Cu, and 5% of As and Pb of the mobile
coastal wetland and other types of environmentally sensitive areas, and                   fraction. Given the simplicity of our integrated tool, this methodology
a wide range of socio-economic situations. Thus, the database would                       could be extremely promising and useful for researchers and managers
arrange an extensive range of variables in order to cover any system                      to use in identifying, characterising, and assessing systems globally
worldwide threatened by mine tailings dams.                                               threatened by mine tailings dams.

7. Conclusions                                                                            Acknowledgements

    Overall, the present study has illustrated the potential of applying a                    This work has been financed by Spanish Ministry of Education and
simple assessment methodology based on combining chemical sequen-                         Science through project CTM2007-66724-C02-02, and by the EU
tial extraction with dam break flood simulation as a more comprehen-                       through project e-EcoRisk. We thank Dr. John G. Farmer (Executive
sive and holistic approach to risk evaluation related to collapse of tailings             Editor) and an anonymous reviewer for their comments and helpful
dams. This methodology was used to assess the magnitude of                                criticisms that significantly improved the quality of the manuscript.
environmental impact that would be caused by the possible collapse
of a roasted sulphide tailings dam located in Almagrera industrial                        References
complex (Iberian Pyrite Belt, southwestern Spain). The tailing dam
breaking up and waste spilling (ca. 4.54 Mt) would cause a high density                   Álvarez-Valero AM, Sáez R, Pérez-López R, Delgado J, Nieto JM. Evaluation of heavy
                                                                                              metal bio-availability from Almagrera pyrite-rich tailings dam (Iberian Pyrite Belt,
mud flow that would cover a total pathway of 80 km in approx. a couple
                                                                                              SW Spain) based on a sequential extraction procedure. J Geochem Explor
of days and affect coastal wetlands with high ecological value (RAMSAR                        2009;102:87–94.
and NATURA 2000 conservation areas). Just considering the mobile                          Andalucia Region Government. Modelo Digital del Terreno de Andalucía: Relieve y
fraction, total quantities of contaminants that could be released into the                    Orografía a resolución 10 m; 2005. http://www.juntadeandalucia.es/medioam-
surrounding environment in the event of entire exposure by a                              Bigham JM, Carlson L, Murad E. Schwertmannite, a new iron oxyhydroxysulphate from
hypothetical stack collapsing are approx. 7.1·104 t of S, 1.6·104 t of Fe,                    Pyhasalmi, Finland, and other localities. Mineral Mag 1994;58:641–8.
                                                   R. Pérez-López et al. / Science of the Total Environment 407 (2009) 5761–5771                                                       5771

Borrego J, Morales JA, Pendon JG. Holocene filling of an Estuarine Lagoon along the          Parker G, Robertson A. Acid Drainage. A critical review of acid generation from sulfide
    Mesotidal Coast of Huelva the Piedras River Mouth, southwestern Spain. J Coastal            oxidation: Processes, treatment and control, Australian Minerals & Energy Environ-
    Res 1993;9:242–54.                                                                          ment Foundation, Occasional Paper no. 11; 1999. 227 pp.
Cánovas CR, Olías M, Nieto JM, Sarmiento AM, Cerón JC. Hydrogeochemical                     Pérez-López R, Álvarez-Valero AM, Nieto JM, Sáez R, Matos JX. Use of sequential
    characteristics of the Tinto and Odiel Rivers (SW Spain). Factors controlling metal         extraction procedure for assessing the environmental impact at regional scale of
    contents. Sci Total Environ 2007;373:363–82.                                                the São Domingos Mine (Iberian Pyrite Belt). Appl Geochem 2008;23:3452–63.
DAMBRK UK User Manual. Department of the Environment, England; 2002. http://                Rauret G, López-Sánchez JF, Sahuquillo A, Rubio R, Davidson C, Ure AM, Quevauviller Ph.
    www.civilsoftware.co.kr/Pds/dambrk_user%20mannual.pdf.                                      Improvement of the BCR three step sequential extraction procedure prior to the
De Saint-Venant B. Theory of Unsteady Water Flow, with Application to River Floods and          certification of new sediment and soil reference materials. J Environ Monitor 1999;1:
    to Propagation of Tides in River Channels. CR Acad Sci Paris 1871;73:237–40.                57–61.
Ferreira RML, Leal JGAB.1D Mathematical modelling of the instantaneous Dam-Break Flood      Rico M, Benito G, Salgueiro AR, Díez-Herrero A, Pereira HG. Reported tailings dam
    Wave Over Mobile Bed: Application of TVD and Flux-Splitting Schemes. Proceedings of         failures: A review of the European incidents in the worldwide context. J Hazard
    Union Européenne, CADAM meeting, Munchen, Germany; 1998. p. 215–27.                         Mater 2008a;152:846–52.
Fread DL. Channel routing. In: Anderson MG, Burt TP, editors. Hydrological Forecasting.     Rico M, Benito G, Díez-Herrero A. Floods from tailings dam failures. J Hazard Mater
    New York: Wiley; 1985.                                                                      2008b;154:79–87.
Funes V, Alhama J, Navas JI, López-Barea J, Peinado J. Ecotoxicological effects of metal    Sáez R, Pascual E, Toscano M, Almodóvar GR. The Iberian type of volcano-sedimentary
    pollution in two mollusc species from the Spanish South Atlantic littoral. Environ          massive sulphide deposits. Miner Deposita 1999;34:549–70.
    Pollut 2006;139:214–23.                                                                 Sahuquillo A, Lopez-Sanchez JF, Rubio R, Rauret G, Thomas RP, Davidson CM, Ure AM.
García de Miguel JM. Mineralogía, paragénesis y sucesión de los sulfuros masivos de la          Use of a certified reference material for extractable trace metals to assess sources of
    Faja Pirítica en el suroeste de la Península Ibérica. Bol Geol Min 1990;101:73-105.         uncertainty in the BCR three-stage sequential extraction procedure. Anal Chim Acta
Gupta SK, Vollmer MK, Krebs R. The importance of mobile, mobilisable and pseudo-total           1999;382:317–27.
    heavy metal fractions in soil for three-level risk assessment and risk management.      Salgueiro AR, Garcia Pereira H, Rico M, Benito G, Díez-Herrero A. Application of
    Sci Total Environ 1996;178:11–20.                                                           Correspondence Analysis in the Assessment of Mine Tailings Dam Breakage Risk in
ICOLD (International Commission on Large Dams. A Guide to Tailings Dams and                     the Mediterranean Region. Risk Anal 2008;28:13–23.
    Impoundments: Design, Construction, Use and Rehabilitation. Bulletin (United            Salomons W. Environmental impact of metals derived from mining activities:
    Nations Environment Programme) no. 106; 1996. 239 pp.                                       processes, predictions, prevention. J Geochem Explor 1995;52:5-23.
ICOLD (International Commission on Large Dams. Tailings dams — risk of dangerous            Sánchez-España J, López Pamo E, Santofimia E, Aduvire O, Reyes J, Barettino D. Acid mine
    occurrences, lessons learnt from practical experiences. Bulletin (United Nations            drainage in the Iberian Pyrite Belt (Odiel river watershed, Huelva, SW Spain):
    Environment Programme) no. 121; 2001. 144 pp.                                               geochemistry, mineralogy and environmental implications. Appl Geochem 2005;20:
Korte F, Spiteller M, Coulston F. The Cyanide Leaching Gold Recovery Process Is a               1320–56.
    Nonsustainable Technology with Unacceptable Impacts on Ecosystems and                   Santos A, Caballero B, Prada JM. Descripción geológica de los yacimientos de Sotiel
    Humans: The Disaster in Romania. Ecotox Environ Safe 2000;46:241–5.                         Coronada. Bol Geol Min 1996;107:511–8.
Lin Z, Qvarfort U. Predicting the mobility of Zn, Fe, Cu, Pb, Cd from roasted sulfide        Simón M, Ortiz I, García I, Fernández E, Fernández J, Dorronsoro C, Aguilar J. Pollution of soils
    (pyrite) residues a case study of wastes from the sulfuric acid industry in Sweden.         by the toxic spill of a pyrite mine (Aznalcóllar, Spain). Sci Total Environ 1999;242:105–15.
    Waste Manage 1996;16:671–81.                                                            Soldán P, Pavonič M, Bouček J, Kokeš J. Baia Mare Accident-Brief Ecotoxicological Report
López-Pamo E, Barettino D, Antón-Pacheco C, Ortiz G, Arránz JC, Gumiel JC, Martínez-            of Czech Experts. Ecotox Environ Safe 2000;49:255–61.
    Pledel B, Aparicio M, Montouto O. The extent of the Aznalcóllar pyritic sludge spill    Ure AM, Quevauviller Ph, Muntau H, Griepink B. Speciation of heavy metal in soils and
    and its effects on soils. Sci Total Environ 1999;242:57–88.                                 sediments. An account of the improvement and harmonisation of extraction
McBride RA. Soil consistency limits. In: Carter MR, editor. Soil Sampling and Methods of        techniques undertaken under the auspices of the BCR of the Commission of the
    Analysis. Boca Raton, FL: Lewis Publications/CRC Press; 1993. p. 519–27.                    European Communities. Int J Environ An Ch 1993;51:135–51.
McCormack RW. Numerical Solution of the Interaction of a Shock Wave with a Laminar          Webster JG, Swedlund PJ, Webster KS. Trace metal adsorption onto an acid mine
    Boundary Layer. Proceedings of the Second International Conference on Numerical             drainage iron(III) oxy hydroxy sulfate. Environ Sci Technol 1998;32:1361–8.
    Method in Fluid Dynamics, Berlin; 1971. p. 151–63.                                      Weisz M, Polyák K, Hlavay J. Fractionation of elements in sediment samples collected in rivers
Monterde, J.R. Ambientes sedimentarios y morfología de la pirita diagenética temprana           and harbors at lake Balaton and its catchment area. Microchem J 2000;67:207–17.
    en los sedimentos recientes de la Ría de Huelva (S.O. España), PhD Thesis,              Younger PL, Banwart SA, Hedin RS. Mine water: Hydrology, Pollution, Remediation.
    University of Huelva, Spain; 2004.                                                          Dordrecht: Kluwer; 2002.
Morales JR. The Sotiel Mine. Mining Mag 1986;154:132–6.                                     Yu KC, Tsai LJ, Chen SH, Ho ST. Correlation analyses binding behavior of heavy metals
Olías M, Nieto JM, Sarmiento AM, Cerón JC, Cánovas CR. Seasonal water quality                   with sediment matrices. Water Res 2001;35:2417–28.
    variations in a river affected by acid mine drainage: the Odiel River (South West
    Spain). Sci Total Environ 2004;333:267–81.
Olías M, Cánovas CR, Nieto JM, Sarmiento AM. Evaluation of the dissolved contaminant
    load transported by the Tinto and Odiel rivers (South West Spain). Appl Geochem

Shared By: