Performance of an open limestone channel for treating a stream affected by acid rock drainage (León, Spain) Esther Santofimia1&Enrique López-Pamo1… [611833]

RESEARCH ARTICLE
Performance of an open limestone channel for treating
a stream affected by acid rock drainage (León, Spain)
Esther Santofimia1&Enrique López-Pamo1
Received: 19 January 2016 /Accepted: 24 March 2016 /Published online: 12 April 2016
#Springer-Verlag Berlin Heidelberg 2016
Abstract The generation of acid rock drainage (ARD) was
observed after the oxidation dissolution of pyrite-rich black
shales, which were excavated during the construction of a
highway in León (Spain). ARDs are characterized by the pres-
ence of high concentrations of sulfate and metals (Al, Fe, Mn,
Zn, Cu, Co, Ni, Th, and U) that affect the La Silva stream.
Dissolved element concentrations showed values between oneand four orders of magnitude higher than those of natural
waters of this area. A passive treatment system was construct-
ed; the aim of which was to improve the quality of the water ofthe stream. This work provides a hydrochemical characteriza-
tion of the La Silva stream after its transit through the different
elements that constitute the passive treatment system (openlimestone channel (OLC), small ponds, and a wetland), during
its first year of operation. The passive treatment system has
two sections separated by a tunnel 230 m long. The first sec-tion, which stretches between the highway and the tunnel
entrance, is an OLC 350 m long with a slope of 16 %. The
second section, which stretches from the tunnel exit to the endwetland, has a length of 700 m and a slope of 6 %; it is in this
section where six small ponds are located. In the first section
of this passive treatment system, the OLC was effectivelyincreasing the pH from 3 to 4 –4.5 and eliminating all of the
dissolved Fe and the partially dissolved Al. These elements,after hydrolysis at a pH 3 –3.5 and 4 –4.5, respectively, had
precipitated as schwertmannite and hydrobasaluminite, whileother dissolved metals were removed totally or partially foradsorption by the precipitates and/or by coprecipitation. Thesecond section receives different inputs of water such as
ARDs and natural waters. After exiting the treatment system,
the stream is buffered by Al at a pH of 4 –4.3, showing high Al
concentrations (19 –101 mg/L) but with a complete removal of
dissolved Fe. Unfortunately, the outflow shows similar orhigher acidity than the inflow into the system due to the dis-
charge of ARDs (mainly from the tunnel) that is received and
to the existence of a natural stream, which is affected by awaste-rock pile. The predictions and calculations necessary
for the design of any remediation/attenuation techniques are
quite difficult. Despite the fact that the selected design is themost adequate one for this valley and type of passive treatment
system (including adequate slopes), we must admit that the
physicochemical characteristics of the ARD were not the mostappropriate according to the literature. Moreover, during the
design, engineers were unaware of the existence of the inflow
from two highly polluting sources, which have rendered thepassive treatment system ineffective and which therefore sug-
gest that certain improvement measures could be considered.
Keywords Open limestone channels .Acid rock drainage .
Passive treatment system .Schwertmannite .
Hydrobasaluminite
Introduction
Acid rock drainage (ARD) is a major pollution problem
throughout the world and one that adversely affects both sur-
face and ground water. ARD occurs when sulfide-rich min-erals are exposed to the weathering effects of oxygen and
water (Nordstrom and Southam 1997 ). Inorganic water pollu-
tion caused by water-sulfide interaction is a globallyResponsible editor: Philippe Garrigues
*Esther Santofimia
e.santofimia@igme.es
1Instituto Geológico y Minero de España (IGME), Ríos Rosas, 23,
28003 Madrid, SpainEnviron Sci Pollut Res (2016) 23:14502 –14517
DOI 10.1007/s11356-016-6562-z

widespread environmental problem. These waters are charac-
terized by high sulfate and metal concentrations (Fe, Al, Cu,
Zn, and Mn in the order of mg/L as the main typical constit-uents in addition to a broad set of minor elements like As, Pb,
Ni, Cd, Cr, or Co among others, in the order of μg/L) and low
pH.
ARDs can be generated by natural processes or anthropo-
genic activities such as airport and highway construction
(Mathews and Morgan 1982 ; Fox et al. 1997 ;O r n d o r f fa n d
Daniels 2004 ; Hammarstrom et al. 2004 ;T o d de ta l . 2007 ).
Although uncommon, the generation of ARDs associated
with road construction is not unusual, especially in areaswhere geological formations contain sulfides. In Virginia
and Pennsylvania (USA), road construction in several
sulfide-rich geological formations have led to the generationof ARDs (Daniels and Orndorff 2003 ; Orndorff and Daniels
2004 ; Hammarstrom et al. 2004 ). In Nova Scotia (Canada),
the excavation of the Halifax Formation (containing abundantpyrrhotite) generated serious problems involving ARDs,
which were particularly severe during the construction works
of the Halifax airport, and to a lesser extent during the con-struction of Highway 107 near Lake Petpeswick (Fox et al.
1997). Moreover, in the province of British Columbia
(Canada), acid drainage generated from the construction of
route 97C resulted in a considerable impact to the Pennask
River trout (Morin and Hutt 2007 ; Walls 2010 ). In Australia,
the problem relates to the excavation of geographically wide-
spread coastal soils containing sulfides, which have been
mapped (Fitzpatrick et al. 2008 ) and for which several evalu-
ation and management guidelines have been prepared (Tulau
2007 ;D e a re ta l . 2002 ).
Polluted acid waters can be remediated via two generic
approaches, active or passive treatment (Johnson andHallberg 2005 ). The former involves the use of energy (e.g.,
for pumping) and the addition of chemicals (e.g., lime and
caustic soda for pH correction), to attenuate water pollutants.
Passive treatment technologies have gained high acceptance
given their lower long-term operating costs and good results.
Many different passive treatment options have been devel-
oped to remediate these waters, the most popular being aero-bic and anaerobic wetlands, compost wetland (Jarvis andYounger 1999 ; Valente et al. 2012 ), anoxic limestone drains
(ALD, Skousen 1991 ; Cravotta and Trahan 1999 ), successive
alkalinity-producing systems (SAPS, Jage et al. 2001 ;M a y e s
et al. 2009), reducing and alkalinity-producing systems
(RAPS, Younger et al. 2002 ), permeable reactive barriers
(PRB, Jarvis et al. 2006 ;G i b e r te ta l . 2011), and dispersed
alkaline substrate (DAS, Rötting 2008 ). OLCs are particularly
useful in steep terrains where long (300 to 1000 m) channelsare possible, offering a unique treatment system where noother passive system is likely to be appropriate
(Ziemkiewicz et al. 1996 ). More often than not, OLCs are
preferred due to their low construction and maintenance costs.These methods are used alone or in combination with other
methods depending on the kind of polluted acid waters and the
characteristics of the site (e.g., space availability, topography).Most of these passive treatments described here were original-
ly designed for underground coal mines.
Unfortunately, some of these systems are prone to clogging
or
loss of reactivity due to the formation of precipitates (metal
oxyhydroxides and hydroxysulfates) when treating water with
high Fe, Al, and other metal concentrations (Cravotta andTrahan 1999 ; Watzlaf et al. 2000 ,2002 ; Ziemkiewicz et al.
2003 ; Watzlaf et al. 2004 ). Other problems include gypsum,
which could precipitate in waters with high SO
4and Ca con-
tent (Hammarstrom et al. 2003 ).
Spain has limited experience with the implementation of
these systems for mitigating acid waters. An example of this isCartagena, where a 1986-m-long limestone channel with a
slope of 4.6 % was constructed, which helped achieve a strong
reduction in Fe and As concentration but not so in the case ofother elements such as SO
4−2, Al, Mn, Zn, Cd, and Pb, whose
concentration exceeded the limits set by the World HealthOrganization by one or two orders of magnitude (Alcoleaet al. 2012), thereby rendering the system somewhat
ineffective.
The main goal of this work is to study the performance of a
passive system for treating acid rock drainage in connectionwith the construction of a highway, and more specifically, as aresult of excavation of pyrite-rich shales. This ARD affect the
La Silva stream, which has suffered from strong environmen-
tal impact.
Materials and methods
Field site
The La Silva stream is located east of the Bierzo region in the
province of León (Spain). It is 7.6 km long and is a tributary of
the Tremor River. The passive treatment system was builtalong a 2-km stretch of the river from the source of the stream
to the town of La Silva.
Passive treatment specifications and design
The La Silva stream reclamation project took part in two main
stages: (i) the removal of up to 100,000 m
3of residual material
coming from the construction of the highway. These are po-
tential generators of ARD that were deposited in an adequate-
ly sealed landfill specifically prepared for avoiding the gener-
ation of acid drainage and (ii) the construction of a passivetreatment system (Figs. 1and2), consisting of a limestone
channel with a width of between 2 and 4 m, six small ponds(Fig. 2a, b ) that act as decanting ponds and a final aerobicEnviron Sci Pollut Res (2016) 23:14502 –14517 14503

wetland (Fig. 2d). A vegetation area with fiber roll was
installed on the right bank of the channel (Fig. 2a).
The treatment device consists of two sections that are
separated by a tunnel 230 m long (Fig. 1). The first section,
between the highway and the entrance to the tunnel, is a
limestone channel 350 m long with a total average slope of
16 % (Fig. 2a, f). The second section, which stretches from
the exit of the tunnel (Fig. 2e) to the final wetland
(Fig. 2d), has a length of 700 m. It is in this section, the
six small decanting ponds are arranged in echelon(Fig. 2b). The average slope in this section is reduced to
6 %. Limestone filter walls w ere built to treat the acid
drainage that flows from the side into the stream throughits right bank (Fig. 2c). Approximately 22,000 m
3of neu-
tralizing material, consisting mainly of limestone, wereused in the construction of this treatment system.Hydraulic network of the treatment system
The built channel along with its interspersed ponds rests on
the very riverbed of the La Silva stream along a 1280-m
stretch. The system is open laterally, as it receives the contri-
bution from several tributaries (La Retuerta, Aborregados, andel Corón), drainage from highway side ditches and diffuse
ARDs coming from the north slope of the valley, which are
generated at the edge of the highway (Fig. 1). Determining the
flow rate of each of the lateral contributions has not been
possible, which makes studying the performance of the treat-
ment system difficult.
Freshwater contributions help to dilute the high concentra-
tions of sulfate and metals of the La Silva stream; in turn,however, contribution from highway side ditches and diffusedrainage from the north slope of the valley are acid drainage
flows that carry contaminant loads.
Water sampling and laboratory analyses
Field measurements and water sampling were carried out be-
tween September 2011 and September 2012. Field parameters
such as pH, redox potential (ORP), temperature (T), dissolved
oxygen (DO), and electric conductivity (EC) were measuredin situ with HANNA multiparametric probe (HI 9828) and
Hydrolab® Quanta probe, properly calibrated on site against
supplied calibration standards solutions pH (4.01 and 7.01),ORP (240 mV), and EC (1413 μS/cm).
Flow rates have been calculated by conventional methods
using digital flow meters (GLOBAL WATER), with the ex-ception LS-3 station, where flow was calculated in a channel
section (vee-shaped).
The concentration of Fe(II) was measured by reflectance
photometry with a Merck RQflex10 reflectometer andReflectoquant® analytical strips. Two different reagents were
used depending on the Fe(II) concentration: (1)Ferrospectral® for the 0.5 –20 mg/L range and (2) 2.2 ′-
bipiridine for the 20 –200 mg/L range. SO
4concentration
(range 40 –150 and 150 –900 mg/L) were measured in situ
through spectrophotometer DR 2800 of Hanch-Lange.
Mineral acidity and alkalinity were also determined by in situ
titration with NaOH 1.6 N (Hach method 8201) and H 2SO4
0.16 N (Hach method 8203). Titration tests for the measure-
ment of AMD acidity were performed at intermediate pH
values of 3.7 and 5, and total acidity (pH 8.3), with NaOH
and continuous measurement of pH and the amount of base
used in every step.
All samples were filtered in situ with 0.45- μmm e m –
brane filters from Millipore, stored in 125-mL polyethylenebottles, acidified with HNO
3, and refrigerated at 4 °C during
transport.
Water samples were analyzed using atomic absorption
spectrophotometry (AAS, Varian SpectrAA 220 FSTunnelLS-1
LS-2
LS-3NW-1
NW-2LS-4
El Corón stream
NW-3
LS-6
LS-7
LS-8Aborregados stream
WetlandHighway A-6
ARD
ARD
ARDARD
La Silva stream (LS)
Natural Water (NW)
Acid drainage from ditch
Pond
ARDAcid Rock DrainageLS-5First section
350 mSecond section
700 m230 m
La Silva Village
Fig. 1 Sketch showing sampling points along the passive treatment
system in La Silva stream14504 Environ Sci Pollut Res (2016) 23:14502 –14517

equipment) for Na, Mg, Ca, Cu, Mn, and Zn; inductively
coupled plasma-atomic emission spectrometry (ICP-AES,
VarianVista MPX equipment) for S, Fe, and Al; and induc-tively coupled plasma-mass spectrometry (ICP-MS, Leco
Renaissance) for Co, Ni, Th, and U. The accuracy of the
analytical methods was verified against certified reference wa-ters (TM-27.3 and TMDA-51.3 of National Water Research
Institute), and close agreement with certified values was
achieved for all metals.
115In was used as an internal standard
for the calibration and measurement of ICP-MS
determinations.
Solid sampling and laboratory analyses
In September 2011 and March 2012, precipitates were sam-
pled in the OLC. After identification and sampling of the solid
phases, all samples were directly stored in 125-mL polyethyl-
ene bottles. Immediately after collection (in laboratory), the
precipitates were repeatedly washed with (MilliQ water),
dried at room temperature, weighed, and stored at ambientconditions.
Solid samples were mineralogically characterized by pow-
der XRD using a PANalytical X ’Pert Pro diffractometer with
Cu Kαradiation (40 kV , 40 mA) and a diffracted-beam mono-
chromator. For routine XRD inspections, 2° –70° 2 θscans
were used with 0.5-s counting time per step.Solid samples were analyzed using XRF (PANalytical
MagiX) for the elements Si, Al, Fe, Ca, Ti, Mn, K, Mg, andP, and elemental analyzer (Eltra CS-800) for total S. Afterdigestion with HF, HClO
4, and HNO 3to dryness and dissolu-
tion with HCl 10 %, V , Cr, Co, Mo, Ag, Cd, Ba, Th, and Uhave been determined with ICP-MS; Cu, Zn, and Pb withAAS; and As, Be, Ni, and Se with ICP-AES. A number of
certified international reference materials (BCS 175/2, BCS
302/1, and BCS 378, from the British Chemical Standards;FER-1 and FER-2 from the Canada Centre for Mineral and
Energy Technology) were used to check the accuracy of the
analytical data.
Geochemical modeling
The Phreeqc geochemical analysis software (Parkhurst and
Appelo 1999 , version 2.18) was used to calculate chemical
equilibrium between the solid and dissolved phases in theOLC and to predict the dominating species and ionic com-plexes in the water as well as the formation of minerals. The
saturation state of water analyses of different stations with
respect to several minerals were calculated using theWateq4f database (Ball and Nordstrom 1991). Additional
thermodynamic data for schwertmannite were taken fromBigham et al. ( 1996 ).
Fig. 2 Different elements of the
constructed passive treatment
system: afirst section, before the
La Retuerta tunnel, consisting of alimestone channel with anaverage slope of 16 %; bsecond
section, after the tunnel, with an
average slope of 6 %, where small
decanting ponds are arranged inechelon along the channel; cfilter
walls for mitigating the impact ofARDs coming from the northernslope of the valley; dfinal
wetland of the treatment system; e
tunnel exit, La Retuerta streamjoins on the left side; and ftunnel
entrance, iron and aluminumprecipitates were observed and
sampledEnviron Sci Pollut Res (2016) 23:14502 –14517 14505

Geological, precipitation, and hydrochemical setting
Geology
The highway built at the source of the La Silva valley rests on
Ordovician black shales (IGME 1973 )w i t hh i g hp y r i t ec o n –
tent that are known locally as Luarca shales. These are char-
acterized as massive, fine-grained gray-black shale formations
with abundant organic matter. Pyrite fills three extensionalfracture systems in the shales (Fernández-García et al. 1984 ).
After the construction of the highway earthworks contrib-
uted to the alteration of pyrite, generating acid drainage withhigh sulfate and metal content, mainly Fe, Al, and Mn
(Villa-Bermejo et al. 2008 ; Vadillo et al. 2009 ;2010a ). A
source area for this type of acid drainage are trench slopes
and stripped zones, as can be deducted from the iron pre-
cipitates that settle in the ditches or from the effects of dis-
solution of aggregates and concrete mortar. The watersflowing through these ditches are collected and driven to
the source of the La Silva stream which is also already
affected by acid drainage, as evidenced by the iron precipi-tates that can be observed in the river bed.
Precipitation
The average annual precipitation during the period 1971 –
2010 was 902 mm (Brañuelas weather station; data from
AEMET (Spain ’s National Weather Service)). For the most
part, the annual precipitation normally remains between 800
and 1030 mm (25 and 75 percentiles, respectively).
Our study period coincides with the 2011/12 hydrologic
year. Precipitation in this period was 543 mm, which is con-siderably below the usual average value of 902 mm. This
makes the 2011/2012 hydrologic year the driest year sincethe Brañuelas weather station began to record data (1971)
and indicates how this study was conducted under exceptional
hydrological conditions.
Hydrochemical characteristics of the La Silva stream
The loss of quality of the La Silva stream due to the construc-
tion of the highway was reflected in the pH and alkalinityvalues that have been recorded since 1998 (Vadillo et al.
2010a ). Previous studies of the hydrochemistry of this stream
upstream from the town of La Silva (Fig. 1) during the 2007 –
2008 period have proven this affection, with acid pH values of3–4 and high sulfate and dissolved metal concentrations in the
following ranges: SO
4−2(1060 –720 mg/L), Fe (8 –24 mg/L),
Al (32 –141 mg/L), Mn (1.3 –3.6 mg/L), Zn (0.5 –15 mg/L), Co
(0.4–1.0 mg/L), Ni (0.6 –1.8 mg/L), and Cu (0.3 –0.8 mg/L)
(Vadillo et al. 2009 ).Results
Water chemistry spatial evolution in the passive treatment
system
The hydrochemical study of the La Silva stream as it flows
through the transit system has been subdivided into two sec-
tions: (1) the first section comprises the stretch running from
its source to the entrance to the tunnel (stations LS-1 and LS-2,respectively; Figs. 1and2), the second section, running from
the exit of the tunnel (LS-3) to the end of the treatment system(LS-8; Fig. 1).
Hydrochemistry of the treatment system ’s first section
Water entering the treatment system (LS-1) is acid (pH
2.5–2.8) and shows high sulfate and metal concentrations
(Table 1). After flowing 350 m through the limestone chan-
nel, its acidity decreases to a pH of 4 –4.5. The highest pH
value along the entire length of the limestone channel is
normally reached at this point (LS-2). The increase in pHresults in dissolved Fe being practically removed in full, in
addition to the partial removal of a series of elements
(Al, Mn, Zn, Cu, Co, and Ni), which is reflected globallyby a decrease in EC between both points (Table 1and
Figs. 3and4). This increase in pH is due to the dissolution
of limestone, as evidenced by the notable increase in Caconcentration.
Since no lateral contributions of water exist in this section,
the reduction in the concentrations due to the effect of disso-lution is ruled out. The effectiveness of the treatment system is
demonstrated by the drop in the total acidity values (Table 1
and Fig. 3).
The tunnel as the main source of pollutants
Water quality deteriorates drastically when the La Silva
stream passes through the La Retuerta tunnel (230 m long).
The concentration of sulfate and metals increases as doesits acidity (Fig. 3and Table 1) as its flow rate increases
considerably. It is evident that the tunnel acts as the maincontaminant source of the system both due to its chemistry
and flow rate. The concentrations increase by an order of
magnitude, and the flow rate multiplies by a factor of 4 –8
(LS-3, Table 1).
It must point out that when the channel was designed,
the section of the tunnel was considered to be chemicallyand hydrologically inert. No variations were to be expected
neither in the chemical composition nor in the volume of
water flowing within it. When becoming aware of the prob-lem after conducting a preinspection of the tunnel
(September 2011), the company responsible for the con-
struction of the treatment system proceeded to remove the14506 Environ Sci Pollut Res (2016) 23:14502 –14517

black shales cobble bed that carpeted the tunnel floor, as it
was considered to be the material with the highest ARD-
generating potential. During this undertaking ARDs, seep-age was observed on the concrete wall. The one that was
most representative in terms of flow rate was sampled. This
ARD showed the highest concentrations of sulfate(3295 mg/L) and metals (Fe 139 mg/L, Al 443 mg/L,
Mn 8.1 mg/L, Ni 5.8 mg/L, Zn 4.4 mg/L, Co 2.8 mg/L,
and Cu 2.4 mg/L) recorded in the area (February 2012data). To attenuate its affection, an alternating sand and
limestone gravel bed was placed in the form of several
dams at the base of the tunnel. Despite all of this, thetunnel continued to behave as the main contaminating
source of the system, as demonstrated by sampling from
December 2011 (Table 1and Figs. 3and4).
Hydrochemistry of the treatment system ’s second section
After the exit of the tunnel, the La Silva stream receives the
contribution from La Retuerta natural stream (NW-1, Table 2
and Fig. 1). The latter has a pH that ranges between neutral
a n ds l i g h t l ya c i d( 4 . 3 –6.4) and low mineralization, as shown
by its EC (44 –153μS/cm). The natural waters in the area
show very low alkalinity values (Table 2), which means theirTable 1 Physicochemical characterization, acidity (measured at pH of 3.7, 5.0, and 8.3), flow and water chemistry of different stations in La Silva
stream through the passive treatment system
Date Station Acidity (mg/L CaCO 3eq.) Q pH CE SO 4Ca Na Mg Fe Al Mn Zn Cu Co Ni Th U
pH 3.7 pH 5.0 pH 8.3 L/s μS/cm mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L mg/L μg/Lμg/L
Sep-11 LS-1 42 202 287 * 2.8 1830 430 46 118.0 28.0 3.94 44.9 2.19 0.49 0.13 0.31 0.63 5.0 7.9
Sep-11 LS-2 10 70 114 * 3.1 1344 298 62 96.0 22.0 0.64 17.9 1.46 0.30 0.09 0.19 0.33 0.5 2.3Sep-11 LS-3 477 1817 2127 * 2.5 3490 2748 67 73.0 115.0 95.3 377.0 8.11 2.96 3.17 2.42 4.66 313.0 149.0Sep-11 LS-4 132 634 804 * 3.0 1710 1040 64 29.0 46.0 17.2 135.0 3.08 1.36 1.03 0.92 1.81 81.0 50.2
Sep-11 LS-6 30 331 427 * 3.1 1226 659 73 20.0 32.0 2.02 76.0 2.29 0.82 0.60 0.59 1.11 26.1 26.9
Sep-11 LS-7 12 214 296 * 3.7 1171 645 110 20.0 30.0 0.52 59.2 2.06 0.71 0.47 0.52 0.99 8.4 19.0Sep-11 LS-8 10 165 231 * 4.0 1264 629 198 20.0 30.0 0.34 19.3 1.83 0.47 0.21 0.40 0.78 0.7 7.3Dec-11 LS-1 108 246 306 4.5 2.9 1844 478 43 21.4 22.4 25.21 37.8 1.34 0.44 0.15 0.59 0.33 5.4 6.1Dec-11 LS-2 0 70 104 * 4.5 1557 510 123 20.0 22.9 0.29 20.5 1.21 0.32 0.05 0.39 0.22 0.3 1.3Dec-11 LS-3 86 484 628 * 3.2 2289 1204 126 19.2 45.8 18.88 102.3 2.78 1.28 0.68 1.64 0.88 74.8 38.9Dec-11 LS-4 6 248 312 * 3.4 1204 583 58 8.2 24.2 8.40 53.2 1.42 0.66 0.36 0.78 0.45 32.4 20.1Dec-11 LS-5 0 198 254 * 4.1 987 516 64 47.4 19.8 1.11 41.0 1.49 0.52 0.27 0.65 0.41 6.7 13.9Dec-11 LS-6 0 158 210 * 4.2 914 446 60 46.3 18.0 0.78 36.8 1.37 0.48 0.25 0.58 0.37 4.6 12.3Dec-11 LS-7 0 158 208 * 4.2 948 470 69 47.9 18.8 0.78 36.2 1.39 0.50 0.25 0.59 0.37 3.4 12.0Dec-11 LS-8 0 134 174 37 4.3 922 448 75 42.2 17.8 0.53 30.4 1.30 0.45 0.22 0.55 0.35 1.9 10.2Mar-12 LS-1 24 234 354 0.2 2.9 2291 593 58 188.0 34.0 4.75 56.9 2.42 0.70 0.13 0.44 0.76 6.6 8.5Mar-12 LS-2 0 58 126 * 4.5 1588 436 100 136.0 27.0 0.39 25.1 1.45 0.39 0.05 0.23 0.42 0.3 1.6Mar-12 LS-3 140 1240 1564 0.8 3.2 3294 2244 87 82.0 105.0 84.20 277.0 6.52 3.09 1.61 2.09 4.00 206.0 117.0Mar-12 LS-4 56 466 600 * 3.6 1502 834 54 37.9 39.0 25.10 102.0 2.48 1.14 0.59 0.76 1.46 63.4 40.4Mar-12 LS-5 28 310 424 * 3.5 1245 643 51 27.9 30.0 11.70 72.6 2.37 0.29 0.44 0.63 1.08 39.1 28.2Mar-12 LS-6 2 284 360 * 3.6 1126 545 49 26.2 26.0 7.54 61.8 2.08 0.79 0.37 0.55 0.94 28.5 23.3Mar-12 LS-7 0 236 298 * 3.9 1051 515 62 25.4 24.7 2.19 52.0 1.97 0.71 0.33 0.50 0.86 15.2 19.5
Mar-12 LS-8 0 220 278 6.2 4.0 1005 537 67 25.2 24.4 0.88 49.0 1.84 0.66 0.30 0.46 0.79 8.2 17.7
Jul-12 LS-1 64 292 344 0.4 2.5 2350 548 54 209.0 30.3 4.65 51.3 2.22 0.67 0.15 0.40 0.78 4.9 6.6Jul-12 LS-2 0 126 160 * 4.0 1810 449 91 189.0 25.2 0.52 28.9 1.43 0.47 0.07 0.27 0.57 0.3 1.3Jul-12 LS-3 208 1172 1492 1.8 2.8 3290 2182 90 135.0 95.4 72.25 256.0 6.22 3.11 1.73 2.04 4.23 132.0 67.7Jul-12 LS-4 84 554 834 * 3.0 2260 1319 69 81.6 57.8 17.56 150.0 3.77 1.82 0.99 1.20 2.47 61.4 38.0Jul-12 LS-5 42 390 580 * 3.1 1640 912 59 52.4 39.8 13.33 97.1 3.10 1.18 0.63 0.85 1.64 36.3 24.4Jul-12 LS-6 70 394 474 * 3.1 1540 807 57 44.5 35.3 9.49 88.6 2.90 1.10 0.57 0.78 1.50 29.4 21.2Jul-12 LS-7 44 328 394 * 3.4 1400 729 70 45.1 33.0 2.36 75.2 2.77 0.99 0.49 0.72 1.34 17.7 18.9Jul-12 LS-8 0 296 362 7.8 3.5 1300 660 72 43.3 31.0 1.14 66.4 2.58 0.88 0.44 0.65 1.24 11.0 16.0Sep-12 LS-1 70 321 409 0.08 2.8 2381 688 63 138.0 39.2 7.94 54.7 3.10 0.81 0.16 0.42 0.88 12.0 12.7Sep-12 LS-2 0 65 115 * 4.6 1708 541 133 128.0 31.5 0.45 23.5 1.62 0.46 0.06 0.25 0.54 0.3 1.6Sep-12 LS-3 227 1438 1822 0.8 3.2 3423 3176 82 76.2 118.0 100.50 284.3 7.27 3.41 1.77 2.27 4.76 198.0 104.0Sep-12 LS-4 88 779 1020 * 3.1 2480 1437 107 58.8 71.9 38.95 155.8 4.14 1.77 0.88 1.21 2.29 88.2 53.8Sep-12 LS-5 35 630 775 * 3.6 2160 1454 137 52.5 66.0 11.70 132.4 4.41 1.56 0.75 1.13 2.27 62.0 45.5Sep-12 LS-6 25 550 710 * 3.3 1985 1206 130 48.4 58.3 7.43 115.8 4.07 1.38 0.64 1.00 2.00 46.3 38.3Sep-12 LS-7 0 410 560 * 3.9 1997 1173 184 49.9 56.6 2.53 97.2 3.99 1.22 0.51 0.87 1.76 22.3 29.2Sep-12 LS-8 0 455 600 6 4.3 2031 1277 199 52.5 64.2 0.83 100.5 4.06 1.30 0.50 0.95 1.95 12.7 28.7
Data from September 2011 to September 2012. Stations 1 and 2 are the first section. Stations from 3 to 8 are the second section
*N o ta n a l y z e dEnviron Sci Pollut Res (2016) 23:14502 –14517 14507

entry into the passive treatment system only helps to improve
the quality of the stream through the process of dilution. Water
from the La Silva stream (LS-4), after receiving the contribu-
tion from La Retuerta stream (NW-1), shows an abrupt drop in
the concentration of practically all of the elements (Table 1
and Figs. 3and4).
Further downstream, the La Silva stream receives another
lateral contribution from the Aborregados stream (NW-2); the
latter has an acid pH (3.4 –4.1, Table 2) and high concentra-
tions of sulfate (119 –313 mg/L), Al (17.2 –39.7 mg/L), and
Mn (1.5 –5.3 mg/L).
As this stream receives ARDs from a black shale waste pile
containing pyrite, the concentration of sulfate and metals in-
creases by two to three orders of magnitude. This contribution
is significant both in terms of its permanent flow rate —despitea very dry year (0.6 –9.8 L/s) —and the Al concentrations,
which provide mineral acidity to the system.
The Corón stream joins downstream from sampling station
LS-5 (Fig. 1), whose hydrochemistry resembles that of the La
Retuerta stream (Table 2), with the exception of the Fe(II) en-
richment, which was detected in September 2011 (8.2 mg/L).Finding dissolved iron in the f orm of Fe(II) in oxygenated wa-
ters (DO 2.5 mg/L) is not frequent , as its oxidation kinetics are
quite fast. The presence of ferrous iron could be explained if itformed some complex with a naturally occurring organic sub-
stance originating from plant de composition (i.e., humic-fulvic
substances, which would slow down its oxidation) (Theis and
Singer 1974;S u z u k ie ta l . 1992; Santana-Casiano et al. 2000).
In addition, the enrichment in dissolved organic matter couldexplain the low DO concent ration found in the stream.0100200300400
0 200 400 600 800 1000 1200 1400 1600Al (mg/L)
0306090120
0 200 400 600 800 1000 1200 1400 1600Fe (mg/L)
0246810
0 200 400 600 800 1000 1200 1400 1600Mn (mg/L)
01234
0 200 400 600 800 1000 1200 1400 1600Zn (mg/L)
Sep-11 Dec-11 Mar-12 Jul-12 Sep-12Distance (m)
Distance (m) Distance (m)Distance (m)2345
0 200 400 600 800 1000 1200 1400 1600pH
Distance (m)
050010001500200025003000
0 200 400 600 800 1000 1200 1400 1600SO4 (mg/L)
Distance (m)05001000150020002500300035004000
0 200 400 600 800 1000 1200 1400 1600EC (μS/cm)
Distance (m)
05001000150020002500
0 200 400 600 800 1000 1200 1400 1600Acicity total (mg/L eq.)
Distance (m)Second section TunnelSecond section Tunnel First sectionFirst section Fig. 3 Spatial and time evolution
of pH, EC, and total acidity values
and of the concentrations of SO 4,
Al, Fe, Mn, and Zn in the La Silvastream (stations LS-1 to LS-8)14508 Environ Sci Pollut Res (2016) 23:14502 –14517

After this last addition, the La Silva stream does not receive
any other contribution with significant volumes of water (LS-
6, LS-7, and LS-8). Small inflows of ARDs exist in this stretch
of the stream (which is active only during rainy seasons) that
originates in the valley slopes (Fig. 1). Flow rates are minimal,
but their sulfate and metal concentrations prevent the quality
of water of the La Silva stream from improving. Thesedrainages are acid (pH 3 –3.5), with high concentrations of
SO42−(700–1400 mg/L) and iron (8 –46 mg/L).
Overall, this section of the passive treatment system up to
the final wetland (Fig. 1)s h o w sas l i g h ti n c r e a s ei np Ha l o n g
with an increase in the concentration of dissolved Ca, which is
a reflection of the ongoing dissolution of limestone that istaking place. This increase in pH helps to fundamentally re-
move Fe and trigger a drop in the concentration of Al, Mn, Zn,
Co, and Ni (Table 1).
Time evolution of the water chemistry in the passivetreatment system
The variability in the flow rate of the hydrological network
hosting the treatment system triggers variations in its
hydrochemistry. Variation in the flow rate of natural water
contributions causes the dilution effect to be nonconstant.Likewise, flow rate variations in the affected contributions
cause the contaminant load reaching the system to be highly
variable.
The hydrochemistry in the first section of the treatment
system (LS-1 and LS-2) has remained relatively constant incontrast to the second section (LS-3 to LS-8), where signifi-
cant variations among the different samples have been ob-
served (Figs. 3and4). In this section, the concentrations of
a good number of elements have increased from the December
2011 campaign to that of September 2012 (Figs. 3and4).
During the study period, the highest pH values along with
the lowest dissolved element concentration values —and01234
0 200 400 600 800 1000 1200 1400 1600Cu (mg/L)
Distance (m)
0123
0 200 400 600 800 1000 1200 1400 1600Co (mg/L)
Distance (m)
012345
0 200 400 600 800 1000 1200 1400 1600Ni (mg/L)
Distance (m)
Sep-11 Dec-11 Mar-12 Jul-12 Sep-12Second section Tunnel First section
Fig. 4 Spatial and time evolution of the concentration of several trace
elements in the La Silva stream (stations LS-1 to LS-8)
Table 2 Field parameters and chemical composition of the natural streams
Date Station Q pH CE HCO 3−Ca Na Mg SO 4 Fe Al Mn Zn Cu Co Ni Th U
L/s μS/cm mg/L mg/L mg/L mg/L mg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L μg/L
Sep-11 NW-1 * 4.3 61 14 6.0 6.0 4.0 9.4 80 80 <d.l. <d.l. <d.l. 1.2 4.8 <d.l. <d.l.
Dec-11 NW-1 * 5.9 70 2 3.2 6.9 1.2 13.3 82 40 58 12 4 5.0 9.0 <d.l. <d.l.Mar-12 NW-1 * 6.4 44 2 3.1 4.3 1.3 8.6 25 20 15 779 1.1 2.1 5.9 <d.l. <d.l.Jul-12 NW-1 * 5.3 70 3 3.6 3.8 1.4 11.7 35 40 16 15 2 4.0 10 <d.l. <d.l.Sep-12 NW-1 * 5.2 153 14 8.2 8.1 3.8 36.5 35 1040 187 78 27 3.0 68 <d.l. 0.36Sep-11 NW-2 * 3.4 286 0 13.0 3.0 18 119 11 17,200 1740 160 96 220 249 0.34 2.78Dec-11 NW-2 9.8 4.0 316 3 6.0 3.0 6.7 166 60 20,900 1500 237 103 271 291 0.30 2.50Mar-12 NW-2 3.6 4.1 389 1 6.9 2.8 8.2 170 57 25,000 1990 291 120 322 373 0.32 2.56Jul-12 NW-2 2 3.7 549 1 1.4 3.1 13.6 299 110 39,700 3540 495 187 577 721 0.59 3.79
Sep-12 NW-2 0.6 3.9 587 0 12.2 4.1 16.7 313 310 36,400 5310 436 108 613 680 0.82 3.68
Sep-11 NW-3 3.2 5.0 52 * 1.1 1.1 0.8 7.5 8220 100 70 <d.l. <d.l. 57 7.9 0.01 0.08Dec-11 NW-3 * 4.8 31 12 1.3 1.3 0.9 8.3 68 213 45 14 1.0 4.0 6.0 <d.l. <d.l.Mar-12 NW-3 * 4.7 63 2 1.9 1.1 1.5 11.0 42 250 755 237 4.4 5.8 10.1 <d.l. <d.l.Jul-12 NW-3 0 4.1 48 3 1.4 47.7 1.3 10.5 40 227 80 17 1.0 6.0 11 <d.l. <d.l.
Data from September 2011 to September 2012
NW-1 La Retuerta stream, NW-2 Aborregados stream, NW-3 El Corón stream, Qwater flow rate, ECelectric conductivity, d.l.detection limit
*Not analyzedEnviron Sci Pollut Res (2016) 23:14502 –14517 14509

therefore lower mineral acidity values —were obtained in
December 2011.
The singular aspect about this campaign is that it was con-
ducted after a period of high precipitation, which means that
the discharge of the La Silva stream as well as that from lateral
tributaries was the highest ever recorded (Table 1and2).
Furthermore, the campaign was also conducted after the
works —which sought to improve the quality of the water —
inside La Retuerta tunnel were completed. On the opposite
end of the spectrum, the campaigns that showed the highest
concentrations were those that were conducted in September
after the low water level. The summer of 2012 was particular-ly dry, in fact, the driest in the last 40 years. Hence, the reason
for such low flow rate records (Tables 1and3), with streams
that were active in September 2011 (i.e., El Corón stream) but
totally dry on this occasion.
The rest of the campaigns showed intermediate values
(March and July 2012). The March campaign took place afteran exceptionally dry period, which favors the increase of dis-
solved elements with respec t to the previous campaign
(December 2011). After this campaign, the month of April
was exceptionally wet, but once again, during the months of
May, June, and July, the values were below average monthly
values. Nonetheless and despite the summer period, flow rates
in the limestone channel during the July campaign slightly
exceed those recorded in the March campaign. Despite thisincrease in flow rate, the concentrations of elements had in-
creased in both campaigns.
From the beginning of the study back in September 2011
and at is conclusion in September 2012, the quality of theeffluent of the treatment system had deteriorated. For exam-
ple, when considering the Al concentration as one of the mostproblematic elements due to its mineral acidity, we can ob-
serve an alarming trend during this period (from 19.3 up to
101 mg/L), with the value of Al concentration having experi-enced a fivefold increase during this time. Given that Al pre-
cipitates only in the last section of the treatment system with a
pH> 4, the increase in its concentration will more likely berelated to the hydrochemical evolution of the system as a whole
rather than to performance losses of the treatment device.
Mineral and chemical characterization of the precipitates
formed in the passive treatment system
Fe and Al precipitate samples were collected after visual iden-
tification in the September 2011 and March 2012 campaigns
(Fig. 2f). The study of Fe precipitates reveals that these consist
basically of schwertmannite. The diffractogram of one of the
samples (Fig. 5a) shows the majority of the bands that are
characteristic of schwertmannite (1.46, 1.51, 1.66, 1.95,
2.28, 2.55, 3.39, and 4.86 Å), as have been described by
Bigham et al. ( 1996)a n dD o l d( 2003). However, its Fe/
S
molar ratio of 3.2, which is below the normal range for this
Table 3 Major (wt%) and minor constituents (mg/kg) of mainly schwertmannite (samples 1 and 2) and hydrobasaluminite (samples 3 and 4), precipitate samples f rom the stations LS-2 and LS-5
Major constituents (wt. %)
Date Samples Mineral SiO 2 Al2O3 Fe2O3 CaO TiO 2 MnO K 2OM g O P 2O5 SO3 PPC
Sep-11 1 Sch. + Gt. + Q + ill 7.13 6.9 52.7 0.16 0.16 <d.l. 0.34 <d.l. 2.60 3.46 30.61
Mar-12 2 Sch. 1.10 2.0 55.0 <d.l. 0.02 <d.l. <d.l. <d.l. 1.96 17.38 22.44Sep-11 3 Hydrob. 0.88 39.6 0.8 <d.l. <d.l. <d.l. <d.l. <d.l. <d.l. 16.65 40.62Mar-12 4 Hydrob. 3.45 38.5 0.8 <d.l. 0.03 <d.l. <d.l. <d.l. <d.l. 35.92 21.23
Minor constituents (mg/kg)Date Samples Mineral Ag As Ba Be Cd Co Cr Cu Mo Ni Pb Se Th U V ZnSep-11 1 Sch. + Gt. + Q + ill 0.12 38.6 52.90 0.45 <d.l. 8.27 95.1 78.9 <d.l. 22.80 7.54 2.82 392 10.70 14.70 34.60
Mar-12 2 Sch. 0.14 25.5 5.29 0.14 <d.l. 0.53 88.1 43.5 0.26 1.89 <d.l. 1.41 338 4.12 4.88 7.11
Sep-11 3 Hydrob. <d.l. 0.9 0.49 2.02 <d.l. <d.l. 10.5 24.0 <d.l. 2.03 2.45 <d.l. 21.4 12.80 0.28 8.81
Mar-12 4 Hydrob. <d.l. <d.l. 2.06 4.11 <d.l. 1.07 15.9 23.7 0.48 2.99 <d.l. <d.l. 4.3 2.63 1.05 7.88
Sch.schwertmannite, Qquartz, illillite, Hydrob. hydrobasaluminite, d.l.detection limit14510 Environ Sci Pollut Res (2016) 23:14502 –14517

mineral (Fe/S molar=4 . 6 –8, e.g., Bigham et al. 1996 ;Y ue ta l .
1999; Kawano and Tomita 2001; Acero et al. 2006;
Santofimia et al. 2015 ), could lead to us to believe that it
contains a certain amount of jarosite, whose ideal Fe/S molar
ratio = 1.5. Schwertmannite is a metastable mineral that tends
to transform into goethite (Bigham et al. 1996 ; Jönsson et al.
2005 ;B u r t o ne ta l . 2007 ; Knorr and Blodau 2007 ; Santofimia
et al. 2015 ), jarosite (Wang et al. 2006 ), or a combination of
both (Acero et al. 2006 ).
The saturation index (SI) of iron minerals calculated using the
Phreeqc code yield positive values (IS > 0) for schwertmanniteand goethite and, to a lower extent, for jarosite, which meansthese might precipitate th ermodynamically (Fig. 6).The other iron precipitate has a Fe/S
molarratio = 15. During
the transformation of schwertmannite into goethite, the release
of the sulfate anion to the solution (Schwertmann and Carlson
2005 ; Jönsson et al. 2005 ) causes a gradual increase of the Fe/
Smolar ratio of the solid phase (Bigham et al. 1996 ). This was
confirmed in the diffractogram, which revealed bothschwertmannite and goethite peaks (Fig. 5b). In addition,
XRD techniques revealed other minerals in this sample, such
as quartz and illite (Fig. 5b), hence the enrichment in Si, K,
and Al in its chemical composition (Table 3, sample 1). The
sampling of Fe and Al precipitates shows certain difficultiesdue to the ease with which the sample becomes contaminated
with other minerals present in the medium. The indicationspoint to the latter event as the reason for the results that were
observed in sample 1, even when schwertmannite is the major
mineral.
As regards the aluminum precipitates and despite that the
interpretation from the diffractograms from these minerals ismore complex than those of iron precipitates due to their even
lower crystallinity, it can be confirmed that the former consists
mainly of hydrobasaluminite, as can be evidenced by the twomost intense reflections observed at 12.6 and 4.58 Å and an
additional band of lower intensity at 2.24 and 1.46 Å (Fig. 5c),
which are typical of this mineral (Tien 1968 ;C l a y t o n 1980 ;
Kim and Kim 2003 ). With respect to its chemical composi-
tion, sample 3 shows an Al content of 39.6 wt.% (as with
Al
2O3) and 16.65 wt.% of S (as with SO 3; Table 3), with an
Al/S molar ratio of 3.7, which is similar to those provided by
other authors for this mineral (Bigham and Nordstrom 2000 ;
Kim and Kim 2003 ). Another of the precipitates sampled0 1 02 03 04 05 06 07 04.58 Å
°2θ Cu Kα2.24 Å 12.6 Å
Hydrobasaluminite1.46Åc
C-1
C-210 20 30 40 50 60 704.86 Å3.34 Å2.50 Å
1.95 Å1.68 Å
1.51 Å
1.45 Å
°2θ Cu KαSchwertmannitea
10 20 30 40 50 60 70
°2θCu Kαb
Sch.Q.
Sch. Sch.ill.
ill.
Gt.ill.
Gt.Gt.
Fig. 5 XRD patterns of aschwertmannite sampled in March 2012; b
schwertmannite and other minerals sampled in September 2011; and c
hydrobasaluminite, for samples taken in March 2012 (C-1) and
September 2011 (C-2)
-12-8-4048
0 200 400 600 800 1000 1200 1400 1600SI
Distance (m)Alunite Basaluminite
Gibbsite Jurbanite-15-10-5051015
0 200 400 600 800 1000 1200 1400 1600SI
Distance (m)Goethite Jarosite-Na
Jarosite-K Schwertmannitea
b
Fig. 6 Saturation index for airon precipitates and baluminum
precipitates. Hydrochemical data correspond to stations where
precipitates were sampled in March 2012Environ Sci Pollut Res (2016) 23:14502 –14517 14511

(sample 4, Table 3) showed an Al content of 38.5 % wt.% (as
Al2O3), a value that approaches the ideal value for this min-
eral, but with a high S content (36 wt.% as SO 3), with a molar
ratio (Al/S = 1.7) that is lower than usual for this mineral (Al/
S = 4). Other authors have obtained lower molar ratios (Al/
S = 2.5) for hydrobasaluminite (Kim and Kim 2003 ); however,
theoretically, this value is closer to the ratio of alunite (Al/
S = 2) or jurbanite (Al/S = 1) (Bigham and Nordstrom 2000 ).
The most probable explanation is that this Al precipitate com-
prises a mixture of several hydroxysulfates that increase the
value of the sulfate in this sample. This observation agrees
with the data calculated with Phreeqc, which predicts a satu-ration of the solution with several mineral phases of Al such as
alunite, jurbanite, and gibbsite (Fig. 6).
Discussion
Mineralogical control of open limestone channelperformance and metal removal
An increase in the pH coupled with a significant reduction of
the acidity of 54 –71 % (Fig. 3) was observed in the first
section of the treatment system. The increase in the pH favors
hydrolysis and the subsequent precipitation of Fe(III).
Furthermore, a decrease in the concentration of Mn, Zn, Cu,
Co, Ni, Th, As, Cr, and U was also noted.
The sampled iron precipitate has shown an enrichment of
the following trace elements: Th > Cr > Cu > As > Zn > Ni >V>U( T a b l e 3), thorium being the most abundant trace ele-
ment, followed by Cr, Cu, and As (Table 3).
A typical characteristic of schwertmannite is its high spe-
cific surface, whereby it is capable of removing dissolvedmetals through adsorption. Schwertmannite has a point of zero
charge at pH
iep7.2 (Jönsson et al. 2005 ) and at low pH has the
property to remain positively charged, meaning it can adsorb
anions or anionic complexes, whereas the opposite occurs at
high pH when it can adsorb metal cations or cationic com-
plexes (Kinniburgh and Jackson 1981;B i g h a ma n d
Nordstrom 2000;R e g e n s p u r ga n dP e i f f e r 2005). The La
Silva stream has consistently shown pH values below 4.5,which means any schwertmannite that may have precipitatedin this medium will be positively charged.
According to the data that was obtained using Wateq4f
databases (Ball and Nordstrom 1991 ) and Minteqa2 (Allison
et al. 1990 ), elements such as As and V show anionic species
as the main aqueous species such as a arsenates (H
2AsO 4−)
and vanadates (VO 2SO4−). In addition and although the dom-
inant species for elements such as Zn and U are Zn(II) andUO
2SO40, respectively, Phreeqc calculations have shown that
both elements can form sulfated anionic species such as
Zn(SO 4)2−and UO 2(SO 4)2−2, which are present in the aque-
ous medium at concentrations of 5 and 12 %, respectively. Theenrichment of Cr (Table 3) can be explained by the replace-
ment of the sulfate anion in the crystalline structure of
schwertmannite by chromate-type oxianions, in addition toarsenate, phosphate, or molybdate (Regenspurg 2002 ). With
respect to other divalent metals such as Cu and Ni, the dom-inant aqueous species are the cations Cu(II) and Ni(II), whichare present at concentrations of 79 and 80 %, respectively.
Coprecipitation processes within schwertmannite can explain
the enrichment in both metals (Burgos et al. 2012 ).
Furthermore, Cu(II) can become incorporated within its crys-
talline structure via substitution of Fe(III) or by superficial
adsorption (Antelo et al. 2013 ).
After the full removal of dissolved Fe, any contribution of
alkalinity will favor the increase in pH up to a value of 4 –4.5,
where it will become stabilized again due to the chemicalbuffer generated by the hydrolysis of aluminum. In this sec-
tion, a loss in the concentrations of dissolved Al between 44 –
60 % was observed (from LS-1 to LS-2, Table 1), from which
it could be concluded that partial hydrolysis and precipitationof this ion had taken place. White Al precipitates were identi-
fied in sampling station LS-2 at the entrance to the tunnel(Fig. 2fand Table 3).
With respect to which mineral phases could precipitate,
Nordstrom and Alpers ( 1999 ) and Bigham and Nordstrom
(2000 ) conclude that aluminum minerals with the greatest
probabilities of forming in sulfated waters are alunite (3 Al(ac) + K
+(ac) + 2 SO 42−(ac) + 6 H 2O↔KAl 3(SO 4)2(OH) 6
(s) + 6 H+(ac)) and hydrobasaluminite (4 Al3++S O=
4+2 2–
46 H 2O↔Al4(SO 4)(OH) 1012–36 H 2O+1 0 H+).
Calculations using Phreeqc have yielded positive SI values
for both minerals at sampling station LS-2 (Fig. 6). The pos-
sible precipitation of alunite has been discarded, as samples do
not show K+values, a cation that is essential to its mineral
structure (Table 3). This mineral can appear in acid environ-
ments but it is linked more to hypersaline lakes (Alpers et al.1992 ) or volcanic regions with evidence of hydrothermal al-
teration (Hemley et al. 1969 ). In addition, it is known that the
majority of aluminum precipitates that are formed in acid wa-ters at these pH values normally have the composition of
hydrobasaluminite (Alpers et al. 1994 ; Chapman et al. 1983 ;
Berger et al. 2000 ; Kim and Kim 2003 ). Hydrobasaluminite is
metastable, basaluminite being more stable than the former.Basaluminite is formed by dehydration of hydrobasaluminite
(Clayton 1980 ;N o r d s t r o m 1982 ). Studies have been conduct-
ed on the stability of basaluminite, which have confirmed that
it also becomes metastable after heating and maturing, and
tends to transform into alunite (Adams and Rawajfih 1977 ;
Nordstrom 1982 ). After studying the results that were obtain-
ed, it can be concluded that our samples point tohydrobasaluminite ( “Mineral and chemical characterization
of the precipitates formed in the passive treatment system ”
section). This mineral also has the capacity to adsorb traceelements, as evidenced by our samples that became enriched14512 Environ Sci Pollut Res (2016) 23:14502 –14517

with Cu, Cr, Zn, Be, Th, Ni, and U (Table 3). Several labora-
tory experiments using pH increments in acid waters have
demonstrated how the precipitation of Al can remove differentpercentages of dissolved elements from the solution (e.g., Cu,
Zn, U, and Ni) (Munk et al. 2002 ; Santofimia 2010 ).
If we shift our focus to the second section of the passive
treatment system (from LS-3 to LS-8), a notable increase inthe concentration of dissolved metals such as Fe, Al, Mn, Zn,
Cu, Co, and Ni can be observed in the La Silva stream once ithas flowed past La Retuerta tunnel. This is reflected in the
degree of mineral acidity, which increases by one order of
magnitude (Table 1and Fig. 3). In this section, the passive
treatment shows a slight increase in pH and a gradual decrease
in the concentration of many of the dissolved elements, due to
both the dilution caused by the inflow of natural waters andthe hydrolysis of Fe. Fe precipitates can be observed in many
zones along the limestone channel.
During the study period a pH in the range of 3.5 –4.3 was
reached at the exit of the passive treatment system (LS-8).Unfortunately, the outgoing effluent from the treatment sys-
tem shows, in the majority of occasions, more acidity than inthe water flowing into it (Table 1). High concentrations of Fe
and Al favor this increase, since the system is subject to sig-nificant inputs of these elements (La Retuerta tunnel and the
Aborregados stream), revealing at several sampling points
concentrations that are higher at the exit of the treatment sys-tem than at its entry (e.g., July and September 2012, Table 1).
This has also been observed in the case of other elements suchas Mn, Zn, Cu, Co, Ni, As, Ba, etc. (Table 1).
Water from the La Silva stream shows an anomaly in the
concentration of Th, with very high values in relation to theconcentrations shown by other natural streams sampled inEurope (0.002 to 0.37 μg/L, with an average value of
0.009 μg/L) (De V os et al. 2006 ). The high values recorded
in sampling station LS-3 must be highlighted, with concentra-tions between 74.8 and 313 μg/L (Table 1). The enrichment of
this element could have its origin in the black shales present inthe area (Ugidos et al. 2004 ), which is usual, as described in
the literature (De Vos et al. 2006 ).
Thorium-enriched systems can also be so in U; this has
been observed in our system with concentrations in the rangeof 38.9 to 149 μg/L (sampling station LS-3, Table 1).
Similarities in properties such as the ionic size and externalelectronic configuration of elements such as Th, U, Ce, and Zrare the reason behind the relationship that exists between their
chemical crystallinity. Th and U have a close relationship and
can be minor constituents of a large number of minerals con-taining rare earths such as monazite (Wedepohl 1978 ).
At low pH, thorium in sulfated waters can be dissolved as a
sulfated ionic complex (Langmuir 1980 ; Mernagh and
Miezitis 2008 ). After running the calculations with Phreeqc
on the chemical speciation of this element, we know that Th ispresent in the solution as Th(SO
4)2at 62 %, and as Th(SO 4)3=at 27 %. Th shows a tendency to be removed from the solution
at low pH (3 –4) via coprecipitation (e.g., Th(SO 4)2·8H 2O,
Mernagh and Miezitis 2008 )o r / a n da d s o r p t i o ni nF ep r e c i p i –
tates (Verplanck et al. 1999 ), which agrees with the high Th
concentrations (392 and 338 mg/kg) shown by Fe precipitates
(schwertmannite, Table 3). These Th concentrations are ex-
ceptional in the chemical composition of this mineral. In ad-
dition, this anomaly has also been recorded in stream sedi-
ments, at concentrations of 262 mg/kg, which is very highwhen compared to other stream sediment samples in the area
(Locutura et al. 2012 ).
Evaluation of OLC as a passive treatment systemand future design considerations
Open limestone channels such as the one built in La Silva are
the simplest and leanest devices for treating acid drainage. It is
a physically robust device that does not become clogged asvertical-flow ponds do, is cheap to build but on the contrary, is
one of the lowest-performing treatment systems and one with
the highest limitations when treating water with high metalsconcentrations.
One of the problems that this type of passive treatment
system poses is the precipitate layer that coats the limestonegravel cover, which causes a reduction of about 50 –80 % in
the dissolution of the limestone (Pearson and McDonnell1975 ; Ziemkiewicz et al. 1994 ). Its optimal performance is
achieved with slopes greater than 12 %, where the flow ve-
locity is such that it keeps precipitates suspended, and where
sediments transported in suspension help keep the limestonesurface free of precipitates (Skousen and Ziemkiewicz 2005 ).
In this sense, the first section of the treatment system in LaSilva has an adequate slope for warranting the performance ofa limestone channel. In this section, an improvement in the
hydrochemistry of the stream is observed after it has flowed
through the OLC, as evidenced by a slight increase in pH (3 –
4.5) and the loss of high percentages in the concentration of
metals such as Fe (83 –99.6 %), Al (43 –60 %), Zn (27 –43 %),
Cu (30 –67 %), Co (33 –46 %), and Ni (27 –47 %) and actinides
such as Th (91 –97.5 %) and U (71.5 –87 %). In this section,
the passive treatment system proved its effectiveness by re-
moving almost all of the Fe as well as aluminum in significantproportions and other elements, which were removed from the
medium fundamentally via the precipitation of Fe and, to a
lesser extent, Al, as described in the previous section.
Furthermore, a loss in acidity in the range of 54 –72 % was
also observed in this section, which corresponds to values that
are far above those provided by Ziemkiewicz et al. ( 1997 )
after studying seven AMDs that had undergone treatmentvia the use of OLCs.
It is known that channels are most effective if they have
high slopes; however, their effectiveness in the reduction ofacidity is also influenced by the length of the channel andEnviron Sci Pollut Res (2016) 23:14502 –14517 14513

residence time (McDonald et al. 2001 ). Our first section has a
length of 300 m and its performance can be considered good
when compared with channels described by Ziemkiewicz andBrant ( 1996 ), which have a length of 400 m, a slope of 8 %
and receive a discharge of 60 L/s, with Fe, Al, and Mn con-centrations of 622, 158, and 49 mg/L, respectively. After thedrainage passes through the channel, a decrease of 66, 35, and
14 % was achieved for Fe, Al, and Mn, respectively.
Nevertheless, Rose and Lourenso ( 2000 )h a v ed e m o n s t r a t e d
that shorter OLCs (127 m) with a slope of only 4.4 % can be
effective.
In the second section of the OLC and despite that the con-
tamination that is contributed by the tunnel and theAborregados stream was not taken into account during the
design of the passive treatment system, the pH had increasedslightly to 3.5 –4.3 and the stream water had lost between 97 –
99.6 % of the dissolved Fe concentration, 65 –95 % of Al and
high percentages of other metal concentrations such as Zn(62–84 %), Cu (68 –94 %), Co (58 –84 %), Ni (59 –83 %), as
well as actinides such as Th (93 –99.7 %) and U (72 –95 %).
Indeed, the concentrations of the contaminant elements arereduced progressively in this second section of the channel,
from LS-3 to LS-8 (Fig. 1), but it is difficult to distinguish how
much of this reduction is due to the effect of dilution of
inflowing natural waters or precipitation in the treatment sys-
tem, as the latter supplies alkalinity. These are the pitfalls ofhaving an open system.
When comparing our data with those obtained by Vadillo
et al. ( 2009 ,2010a ,b) during 2008 and 2009 at a specific
location in the La Silva stream near sampling station LS-8,
the treatment was considered to be effective with respect to the
concentration of Fe, with values of less than 1 mg/L at the exitof the limestone channel. In the case of Al, the range of dis-
solved concentrations found in our study (19 –101 mg/L) is
not too far off from that which was obtained before the con-
struction of the treatment system (32 –141 mg/L); neverthe-
less, when considering its current average concentration(53 mg/L) and the previous concentration (83 mg/L), it isbelieved that a certain improvement had taken place, but in
this case, it must be more closely associated with the removal
of acid-generating waste than with the effectiveness of thetreatment system itself.
With respect to the suitability and effectiveness of the dif-
ferent passive treatment systems, Cravotta ( 2010 ), after study-
ing numerous types of passive treatment systems, concludes
that limestone channels are the ones with the lowest perfor-
mance and the highest cost per ton of treated acidity.Generally, this type of treatment is used for slightly acid drain-
age with a relatively low concentration of Fe and Al (<2 mg/
L), which are characteristics that are far removed from the aciddrainage of the La Silva stream in its first section and even
more so in the second. Along these lines, Hammarstrom et al.
(2003 ) consider that a limestone channel is not an advisableoption for treating very acid waters that have a high Fe con-
tent, due to the coating of the limestone by the precipitates.
During the study period, Fe precipitates have been present inthe channel (Fig. 2) but even so, an increase in the concentra-
tion of dissolved Ca was observed, which evidences the dis-solution of limestone (Table 1).
With respect to the adaptation of this passive treatment
system in the La Silva valley, it must be noted that the formeris most likely the only option possible given the topography ofthe valley. The latter is narrow, which means the stream is
quite boxed in. The treatment device built occupies the bottom
of the valley practically along its full length. The limestone/ponds ensemble that was built was the only possible option
given the characteristics, as no space was available for other
types of passive treatment systems that could occupy a largersurface. In addition, the steep slope of the valley is ideal for
preserving the reactivity capacity of the limestone for longer
periods of time. It must be accepted however that performancemight be quite limited due to the hydrochemical characteris-
tics of the stream, which are far removed from those recom-
mended in the literature.
After the study of the data obtained between 2011 and
20
12, some improvements could be considered in the original
design of this passive treatment system. As demonstrated, the
La Retuerta tunnel is one of the most problematic sources and
consequently, a tank that could provide alkalinity could bebuilt at its entry. This suggestion could be viable since it would
not involve high costs and the site would be suitable, both in
terms of space and safety. The proposal would entail combin-ing a passive treatment system (OLC) with an active one,
therefore supplying additional alkalinity (e.g., calcium hy-
droxide (slaked lime), sodium hydroxide (caustic soda) orcaustic magnesia (MgO)) to attempt to neutralize the strong
inflow of acidity that is favored by the increase in dissolved Fe
and Al. In order to evaluate the economic viability of thisproposal, a number of calculations of the quantity of alkaline
material that would be needed to counter the acidity can be
made. When using the value of acidity (1526 mg/L eq.CaCO
3) and the average flow rate ( ∼1 L/s) of station LS-3,
we know we would need 1068 mg/L of Ca(OH) 2and
1221 mg/L of Na(OH) 2, which translates to 92 kg/day of
Ca(OH) 2and 105 kg/day of Na(OH) 2.
Conclusions
Acid rock drainage is generated in the La Silva valley as aresult of the oxidation of pyrite present in black shales that
were excavated during the construction of the A-6 highway in
the town of Villagatón, León province (Spain).
Acid drainage is characterized by high concentrations of
sulfate and a series of metals (Al, Fe, Mn, Zn, Cu, Co, Ni,Th, and U). The concentration of these elements is between14514 Environ Sci Pollut Res (2016) 23:14502 –14517

one and four orders of magnitude higher than that of surround-
ing natural waters. Water from unaffected courses has a low
mineralization (very low EC) and very low alkalinity. Thislatter characteristic does not favor the increase in pH of acid
drainage when both waters become mixed, which is necessary
for the precipitation of the dissolved metals.
In order to counter this affection, a passive treatment sys-
tem was built which consisted of an open channel, smallecheloned decanting ponds, and a final wetland. The latterensemble was considered to be the only viable option due to
the topography of the valley. Although the slopes present at
t h es i t eh a v ef a v o r e dt h ec o n s t r u c t i o no fa nO L C ,w em u s tadmit that according to the literature, the physicochemical
characteristics of the ARD were not the most adequate. In
turn, our passive treatment system is an open one, as it re-ceives numerous contributions of acid drainage and natural
waters. The latter coupled with the fact that no gauging sta-
tions could be installed to measure the flow made it impossibleto run calculations of the performance of the treatment system,
but did enable a follow-up on the evolution of contaminant
element concentration and pH in the system.
A high degree of effectiveness of the OLC was observed in
the first section of the passive treatment system —which does
not receive any inflows of natural water —as evidenced by the
almost complete removal of all of the Fe and a high percentageof Al and other elements (Zn, Cu, Co, Ni, Th, and U). Thelatter were removed from the medium fundamentally via the
precipitation of the Fe (schwertmannite) and to a lesser extent
of aluminum (hydrobasaluminite), resulting in a loss of be-tween 54 and 72 % acidity. With respect to the second section
of the OLC, the stream receives inflows of both acid drainage
and natural water. Prior to the construction of this treatmentsystem, no knowledge was available about highly contami-
nant sources such as (1) the La Retuerta tunnel and (2) the
Aborregados stream, which did not contribute to improvingthe quality of the stream as it flowed through the OLC.
Despite this and the particularly dry year in which the study
was conducted, the passive treatment system achieves, in themajority of the cases, the total removal of Fe. In almost all
occasions, the La Silva stream leaves the system buffered by
the aluminum with a pH between 4 and 4.3. This means thatthe majority of Al has remained dissolved. Indeed, the Fe
concentration at this point has been generally <1 mg/L, where-
as the concentration of Al has remained in the range of 19 –
101 mg/L.
During the study, a strong anomaly was observed in the
dissolved concentrations of Th, which yielded very highvalues (75 –313μg/L). In addition, it has been confirmed that
the schwertmannite precipitates sampled in the stream areenriched in this element, which is exceptional (between 338and 392 mg/kg).
Despite certain improvements, unfortunately, water
flowing through the passive treatment system comes out withvery high acidity values, in contrast to the values recorded at
the entry of the system. The reason is the main contaminant
source, which is the acid drainage inside the tunnel.
When taking into consideration the concentrations of Fe
and Al at the exit of La Retuerta tunnel, the alkalinity thatmust be supplied in order to trigger the precipitation of Al isthree to five times higher than that which is necessary for the
precipitation of Fe. Therefore, Al is the main problem when it
comes to increasing the pH to normal values. Other dissolvedmetals such as Cu, Zn, Co, Ni, and Mn, require higher pH
values for precipitation, which is why they remain mostly
dissolved. Nevertheless, there is a certain amount of theseelements that is removed from the solution due to adsorption
and coprecipitation phenomena taking place in Fe precipitates.
The chemical analyses conducted on these precipitates revealand enrichment of a broad set of these elements.
After an analysis of the data that was collected during this
study several improvements could be suggested for the origi-nal design of this passive treatment system; the latter could be
combined with an active system (e.g., a vessel for the supply
of alkalinity) to counter or attempt to mitigate one of the mostproblematic pollutant sources: the ARDs entering through the
La Retuerta tunnel.
Acknowledgments This work has been supported with funds from
IGME and TRAGSA. We acknowledge the support provided by Jesús
Reyes during laboratory work and TRAGSA during fieldwork. M. Isabel
Prudêncio and an anonymous reviewer are thanked for their helpful com-
ments of this manuscript.
References
Acero P, Ayora C, Torrentó C, Nieto JM (2006) The behavior of trace
elements during schwertmannite precipitation and subsequent trans-formation into goethite and jarosite. Geochim Cosmochim Ac 70:4130 –4139
Adams F, Rawajfih Z (1977) Basaluminite and alunite: a possible cause
of sulphate retention by acid soils. Soil Sci Soc Am J 41:686 –692
Alcolea A, Vázquez M, Caparrós A, Ibarra I, García C, Linares R,
Rodríguez R (2012) Heavy metal removal of intermittent acid mine
drainage with an open limestone channel. Miner Eng 26:86 –98
Allison JD, Brown DS, Novo-Gr adac KL (1990) MINTEQA2/
PRODEFA2, A Geochemical Assessment Model for Environmental
Systems, Version 3.0 User ’s Manual. Environmental Research
Laboratory. Office of Research and Development, U.S. EPA, Athens
Alpers CN, Rye RO, Nordstrom DK, White LD, King BS (1992)
Chemical, crystallographic, and isotopic properties of alunite and
jarosite from acid hypersaline Australian lake. Chem Geol 96:203 –
226
Alpers CN, Blowes DW, Nordstrom DK, Jambor JL (1994) Secondary
minerals and acid mine-water chemistry. In: Jambor JL, Blowes DW
(eds) Environmental Geochemistry of Sulfide Mine-Wastes,
Mineral. Assoc. Canada Short Course Handbook, vol 22., pp 247 –
270
Antelo J, Fiol S, Gondar D, Pérez C, López R, Arce F (2013) Cu(II)
incorporation to schwertmannite: effect on stability and reactivity
under AMD conditions. Geochim Cosmochim Ac 119:149 –163Environ Sci Pollut Res (2016) 23:14502 –14517 14515

Ball JW, Nordstrom DK (1991) User ’s manual for WATEQ4F with re-
vised thermodynamic database and test cases for calculating speci-
ation of major, trace and redox elements in natural waters.Geological Survey Water-Resources Investigation, U.S, pp 91 –
183, Report
Berger AC, Bethke CM, Krumhans JL (2000) A process model of natural
attenuation in drainage from a historic mining district. Appl
Geochem 15:655 –666
Bigham JM. and Nordstrom DK (2000) Iron and aluminium
hydroxysulfates from acid sulphate waters. In: Alpers CN, Jambor
JL and Nordstrom DK (eds) Sulfate minerals: crystallography, geo-chemistry, and environmental significance. Reviews in mineralogy& geochemistry, 40. Mineralogical Society of America and the
Geochemical Society. Chantilly, Virginia, USA, pp 351 –403
Bigham JM, Schwertmann U, Traina SJ, Winland RL, Wolf M (1996)
Schwertmannite and the chemical modeling of iron in acid sulfate
waters. Geochim Cosmochim Ac 60:2111 –2121
Burgos WD, Borch T, Troyer LD, Luan F, Larson LN, Brown J, Lambson
J, Shimizu M (2012) Schwertmannite and Fe oxides formed bybiological low pH Fe(II) oxidation versus abiotic neutralization:
impact on trace metal sequestration. Geochim Cosmochim Ac 76:
29–44
Burton ED, Bush RT, Sullivan LA, Mitchell DRG (2007) Reductive
transformation of iron and sulfur in schwertmannite-rich accumula-tions associated with acidified coastal lowlands. GeochimCosmochim Ac 71:4456 –4473
Chapman BM, Jones DR, Jung RF (1983) Processes controlling metal ion
attenuation in acid mine drainage streams. Geochim Cosmochim Ac
47:1957 –1973
Clayton T (1980) Hydrobasaluminite and basaluminite from Chickerell,
Dorset. Mineral Mag 43:931 –937
Cravotta CA (2010) Abandoned mine drainage in the Swatara Creek
Basis, Southern Anthracite Co alfield, Pennsylvania, USA: 2.
Performance of treatment systems. Mine Water Environ 29(3):
200–216
Cravotta CA, Trahan MK (1999) Limestone drains to increase pH and
remove dissolved metals from acidic mine drainage. Appl Geochem
14:581 –606
Daniels WL, Orndorff ZW (2003) Acid rock drainage from highway and
construction activities in Virginia, USA. Proceedings 6th ICARD,
Cairns, North Queensland, Australia, 12 –18 July 2003, p. 479 –487
De V os W, Tarvainen T et al. (2006) Geochemical atlas of Europe. Part
2—interpretation of geochemical maps, additional tables, figures,
maps, and related publications. Geological Survey of Finland,Otamedia Oy, Espoo, p 692
Dear SE, Moore NG, Dobos SK, Watling KM, Ahern CR (2002) Soil
management guidelines. In: Queensland Acid Sulfate TechnicalManual version 3.8. Department of Natural Resources and Mines,Indooroopilly
Dold B (2003) Dissolution kinetics of schwertmannite and ferrihydrite in
oxidized mine samples and their detection by differential X-ray dif-fraction (DXRD). Appl Geochem 18:1531 –1540
Fernández-García L, Moro-Gómez C, Gómez-Prieto JA, Alvarez del
Campo C (1984) Revisión y síntesis Geológico-Minera de la
Cuenca carbonífera de. IGME, El Bierzo, p 98
Fitzpatrick R, Marvanek S, Powel B (2008) Atlas of Australian acid sul-
phate soils. In: Inland Acid Sulphate Soils System Across Australia.
In: Fitzattrick R, Shand P (eds) CRC LEME Open File Report No.249. (Thematic Volume) CRC LEME, Perth, Australia, pp 75 –89
Fox D, Robinson C, Zentilli M (1997) Pyrrhotite and associated sulphides
and their relationship to acid rock drainage in the Halifax Formation,
Meguma Group, Nova Scotia. Atlantic Geol 33:87 –103
Gibert O, Rötting T, Cortina JL, de Pablo J, Ayora C, Carrera J, Bolzicco J
(2011) In-situ remediation of acid mine drainage using a permeable
reactive barrier in Aznalcóllar (SW Spain). J Hazard Mater 191:287–95Hammarstrom JM, Sibrell PL, Belkin HE (2003) Characterization of
limestone reacted with acid-mine drainage in a pulsed limestonebed treatment system at the Friendship Hill National HistoricalSite, Pennsylvania, USA. Appl Geochem 18:1705 –1721
Hammarstrom JM, Brady K, Cravotta CA (2004) Acid-rock drainage at
Skytop, Center County, Pennsylvania, 2004. U.S. GeologicalSurvey Open-File Report 2005 –1148, pp. 37
Hemley JJ, Hosterler PB, Gude AJ, Mountjoy WT (1969) Some stability
relations of alunite. Econ Geol 64:599 –61
2
IGME (1973) Mapa geológico de España. Hoja 159. Bembribe
Jage CR, Zipper CE, Noble R (2001) Factors affecting alkalinity gener-
ation by successive alkalinity-producing systems: regression analy-
sis. J Environ Qual 30(3):1015 –1022
Jarvis AP, Younger PL (1999) Design construction and performance of a
full-scale wetland for mine spoi l drainage treatment, Quaking
Houses, UK. J Chart Inst Water Environ Manage 13:313 –318
Jarvis AP, Moustafa M, Orme PHA, Younger PL (2006) Effective reme-
diation of grossly polluted acidic, and metal-rich, spoil heap drain-age using a novel, low-cost, p ermeable reactive barrier in
Northumberland, U.K. Environ Pollut 143:261 –268
Johnson D, Hallberg K (2005) Acid mine drainage remediation options: a
review. Sci Total Environ 338:3 –14
Jönsson J, Persson P, Sjöberg S, Lövgren L (2005) Schwertmannite pre-
cipitated from acid mine drainage: phase transformation, sulphaterelease and surface properties. Appl Geochem 20:179 –191
Kawano M, Tomita K (2001) Geochemical modeling of bacterially in-
duced mineralization of schwertmannite and jarosite in sulfuric acidspring water. Am Mineral 86:1156 –1165
Kim JJ, Kim SJ (2003) Environmental, mineralogical, and genetic char-
acterization of ochreous and white precipitates from acid mine drain-ages in Taebaeg, Korea. Environ Sci Technol 37:2120 –2126
Kinniburgh DG, Jackson ML (1981) Cation adsorption by hydrous metal
oxides and clay. In: Anderson MA, Rubin AJ (eds) Adsorption of
inorganic at Solid –liquid interfaces. Ann Arbor Science, Ann Arbor,
Michigan, pp 91 –160
Knorr KH, Blodau C (2007) Controls on schwertmannite transformation
rates and products. Appl Geochem 22:2006 –2015
Langmuir D (1980) The mobility of thorium in natural waters at low
temperatures. Geochim Cosmochim Ac 44:1753 –1766
Locutura J, Bel-Lan A, García Cortés A, Martínez S (2012) Geochemical
Atlas of Spain (Atlas geoquímico de España). Instituto Geológico y
Minero de España, Madrid, p 592
Mathews RC Jr, Morgan EL (1982) Toxicity of Anakeesta Formation
leachates to shovel-nosed salamander, Great Smoky MountainsNational Park. J Environ Qual 11(1):102 –106
Mayes WM, Batty LC, Younger PL, Jarvis AP, Kõiv M, V ohla C, Mander
U (2009) Wetland treatment at extremes of pH: a review. Sci TotalEnviron 407(13):3944 –3957
McDonald L, Sun Q, Skousen J, Ziemkiewicz P (2001) Water quality
prediction model in open limestone systems. In: Xie, Wang, Jiang(eds.) Computer applications in the mineral industries, Swets &
Zeitlinger, Lisse, pp 279 –282
Mernagh TP, Miezitis Y (2008) A review of the geochemical processes
controlling the distribution of thorium in the earth ’sc r u s ta n d
Australia ’s thorium resources. Geoscience Australia Record
2008/05, 48p
Morin KA, Hutt NM (2007) Highway 97C toad-cut environmental pros-
ecution near Pennask Creek. MDAG.com Internet Case Study 27.
Available from http://www.mdga.com/case_studies/cs27.html
Munk LA, Faure G, Pride DE, Bigham JM (2002) Sorption of trace
metals to an aluminum precipitate in a stream receiving acid rock-
drainage; Snake River, Summit County, Colorado. Appl Geochem
17(4):421 –430
Nordstrom DK (1982) The effect of sulphate on aluminum concentrations
in natural waters: some stability relations in the system Al
2O3-SO 3-
H2O at 298° K. Geochim Cosmochim Ac 46:681 –69214516 Environ Sci Pollut Res (2016) 23:14502 –14517

Nordstrom DK, Alpers CN (1999) Geochemistry of acid mine waters. In:
Plumlee GS, Logsdon MJ (eds) The environmental geochemistry of
mineral deposits. Reviews in Economic Geology, vol 6A. SEG,Littleton, CO, USA, pp 133 –156
Nordstrom DK, Southam G (1997) Geomicrobiology of sulphide mineral
oxidation. In: Banfield JF, Nealson KH (eds) Geomicrobiology:Interactions between microbes and minerals, vol 35. MineralogicalSociety of America, Washington, pp 361 –390
Orndorff ZW, Daniels WL (2004) Evaluation of acid-producing sulfidic
materials in Virginia highway corridors. Environ Geol 46:209 –216
Parkhurst DL, Appelo CAJ (1999) User ’s guide to Phreeqc (version 2 —a
computer program for speciation, batch-reaction, one-dimensionaltransport, and inverse geochemical calculations. U.S. Geological
Survey Water-Resources Investigations, Denver, pp 99 –4259
Pearson FH, McDonnell AJ (1975) Use of crushed limestone to neutralize
acid wastes. J Environ Eng Division-ASCE 101:139 –158
Regenspurg S (2002) Characterisation of schwertmannite-geochemical
interactions with arsenate and chromate and significance in sedi-
ments of lignite opencast lakes. pH. D. dissertation. UniversityBayreuth, Bayreuth
Regenspurg S, Peiffer S (2005) Arsenate and chromate incorporation in
schwertmannite. Appl Geochem 20:1226 –1239
Rose AW, Lourenso FJ (2000) Evaluation of two open limestone chan-
nels for treating acid mine drainage. in Proceedings of the 17thAnnual National Meeting of the American Society for Surface
Mining and Reclamation, Florida, pp. 236 –247.
Rötting TS, Caraballo MA, Serrano JA, Ayora C, Carrera C (2008) Field
application of calcite Dispersed Alkaline Substrate (calcite-DAS) for
passive treatment of acid mine drainage with high Al and metal
concentrations. Appl Geochem 23:1660 –1674
Santana-Casiano JM, González-Dávila M, Rodríguez MJ, Millero FJ
(2000) The effect of organic compounds in the oxidation kinetics
of Fe(II). Mar Chem 70:211 –222
Santofimia E (2010) Evolución hidroquímica del lago minero de
Aznalcóllar, Sevilla. pH. D (Thesis). Facultad de CC. Geológicas.Universidad Complutense de Madrid, pp. 303
Santofimia E, López-Pamo E, Montero E (2015) Selective precipitation
of schwertmannite in a stratified acidic pit lake of Iberian Pyrite Belt.Mineral Mag 79(2):497 –513
Schwertmann U, Carlson L (2005) The pH-dependent transformation of
schwertmannite to goethite at 25°C. Clay Miner 40:63 –66
Skousen J (1991) Anoxic limestone drains for acid mine drainage treat-
ment. Green Lands 21(4):30 –35, Charleston
Skousen J, Ziemkiewicz P (2005) Performance of 116 passive treatment
systems for acid mine drainage . In: National Meeting of the
American Society of Mining and Reclamation, Breckenridge,Colorado, June 19 –23. Published by ASMR, 3134 Montavesta
Rd., Lexington, KY 40502, pp. 1100 –1133
Suzuki Y , Kuma K, Kudo I, Hasebe K, Matsunaga K (1992) Existence of
stable Fe(II) complex in oxic river water and its determination.Water Res 26(11):1421 –1424
Theis TL, Singer PC (1974) Complexation of iron(II) by organic matter
and its effect on iron(II) oxygenation. Environ Sci and Technol 8(6):569–573
Tien PL (1968) Hydrobasaluminite and basaluminite in Cabaniss
Formation (Middle Pennsylvani an), Southeastern Kansas. Am
Mineral 53:722 –732
Todd AS, McKnight DM, Jaros CL, Marchitto TM (2007) Effects of acid
rock drainage on stocked rainbow trout (Oncorhynchus mykiss): anin-situ, caged fish experiment. Environ Monit Assess 130:111 –127
Tulau MJ (2007) Acid sulfate soils remediation guidelines for coastal
floodplains in New South Wales. Department of Environment andClimate Change, pp. 136
Ugidos JM, Barba P, Lombardero M (2004) Caracterización geoquímica
de las pizarras negras de las formaciones del Ordovícico Medio-Silúrico del sinclinal de Truchas (Geochemical characterization ofblack shales in the Middle Ordovician-Silurian formations of the
Truchas syncline). Geogaceta 36:27 –29
V
adillo L, Aduvire O, Alberruche E, Arranz JC, Lacal M, Martínez B,
Moreno C, de la Losa A (2009) Evaluación del impacto y
restauración ambiental de la cuenca minera de El Bierzo.
Publicación del Instituto Geológico y Minero de España, Informes
Técnicos n° 1, pp 1571
Vadillo L, Rodríguez V , Alberruche E, Aduvire O, Lacal M, Monteserín
V , Iribarren I, (2010a) Environmental impacts of the A6 Highway in
La Silva Stream in the region of Bierzo. (León). Spain. In: Rauch S
et al. (eds.), Highway and Urban Environment, Alliance for GlobalSustainability Book Series 17:275 –283
Vadillo L, Alberruche E, Rodríguez V, Herrero T, Lacal M (2010b)
Impacto ambiental de la minería del carbón en sistemas acuáticos.Informe interno del IGME, pp 491
Valente TM, Antunes M, Sequeira Braga MA, Prudêncio MI, Marques R,
Pamplona J (2012) Mineralogical attenuation for metallic remedia-
tion in a passive system for mine water treatment. Environ Earth Sci
66(1):39 –54
Verplanck PL, Nordstrom DK, Taylor HE (1999) Overview of rare earth
element investigations in acid waters of US Geological Survey aban-
doned mine lands watersheds. US Geol Surv Wat-Resour Invest Rep
99(4018A):83 –92
Villa-Bermejo E, Alcalde-Aparicio S, Alonso-Herrero E, Vidal-Bardán M
(2008) Presencia de alunita y otros sulfatos de aluminio en los
drenajes ácidos de la minería de carbón del Bierzo (León). Macla
10:147 –148
Walls LD, Li LY , Hall KJ (2013) Trace metal contamination due to acid
rock drainage and its impacts on the fish-bearing Pennask Creek
Watershed, British Columbia. Water Air Soil Pollut 224:1675
Wang H, Bigham JM, Tuovinen OH (2006) Formation of schwertmannite
and its transformation to jarosite in the presence of acidophilic iron-
oxidizing microorganisms. Mater Sci Eng C 26:588 –592
Watzlaf GR, Schroeder KT, Kairies CL (2000) Long term performance of
anoxic limestone drains for the treatment of coal mine drainage.Mine Water Environ 19:98 –110
Watzlaf GR, Kairies CL, Schroeder KT, Danehy T, Beam R (2002)
Quantitative results from the flushing of four reducing andalkalinity-producing systems. In: 2002 West Virginia Surface MineDrainage Task Force Symp, Morgantown, WV
Watzlaf GR, Schroeder KT, Kleinmann RLP, Kairies CL, Nairn RW
(2004) The passive treatment of coal mine drainage. LaboratoryReport DOE/NETL-2004/1202. US Department of Energy,National Energy Technology, Pittsburgh
Wedepohl KK (Exec. Ed) (1978) Handbook of Geochemistry. V olume 2,
Part 5. 266 figs., 576 tables, xxxii + 1546 pages. Berlin, Heidelberg,New York: Springer-Verlag
Younger PL, Banwart SA, Hedin RS (2002) Mine Water: Hydrology,
Pollution, Remediation. Kluwer Academic Publishers, Dordrecht
Yu JY , Heo B, Choi IK, Cho JP, Chang HW (1999) Apparent solubilities
of schwertmannite and ferrihydrite in natural stream water pollutedby mine drainage. Geochim Cosmochim Ac 63:3407 –3416
Ziemkiewicz PF, Brant DL (1996) The Casselman River Restoration
Project. In: Proceedings Eighteenth West Virginia Surface MineDrainage Task Force Symposium 1996, Morgantown, West Virginia
Ziemkiewicz PF, Skousen JG, Lovett R (1994) Open limestone channels
for treating acid mine drainage: a new look at an old idea. GreenLands 24(4):36 –41
Ziemkiewicz P, Brant DL, Skousen J (1996) Acid mine drainage treat-
ment with open limestone channels. Spring, Green lands, pp 33 –40
Ziemkiewicz PF, Skousen JG, Brant DL, Sterner PL, Lovett RJ (1997)
Acid mine drainage treatment with armored limestone in open lime-stone channels. J Environ Qual 26:560 –569
Ziemkiewicz PF, Skousen JG, Simmons J (2003) Long-term performance
o fp a s s i v ea c i dm i n ed r a i n a g et r e a t m e n ts y s t e m s .M i n eW a t e rEnviron 22:118 –129Environ Sci Pollut Res (2016) 23:14502 –14517 14517