Cerro de pasco

Applied Geochemistry 25 (2010) 11071119

Contents lists available at ScienceDirect

Applied Geochemistry

journal homepage: www.elsevier .com/ locate/apgeochem

Geochemistry of highly acidic mine water following disposal into a natural lakewith carbonate bedrock

Christian Wisskirchen a,*, Bernhard Dold a,b, Kurt Friese c, Jorge E. Spangenberg a, Peter Morgenstern d,Walter Glaesser e

a Institute of Mineralogy and Geochemistry, University of Lausanne, CH-1015 Lausanne, Switzerlandb Instituto de Geologa Economica Aplicada, Universidad de Concepcin, Concepcin, Chilec UFZ Helmholtz Centre for Environmental Research, Department of Lake Research, D-39114 Magdeburg, Germanyd UFZ Helmholtz Centre for Environmental Research, Department of Analytical Chemistry, D-04318 Leipzig, Germanye Institute of Geophysics and Geology, University of Leipzig, D-04211 Leipzig, Germany

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

Article history:Received 21 November 2008Accepted 26 April 2010Available online 29 April 2010

Editorial handling by B. Wang

0883-2927/$ - see front matter 2010 Elsevier Ltd. Adoi:10.1016/j.apgeochem.2010.04.015

* Corresponding author. Present address: GoldeSeptiembre 2353 Piso 2, Providencia, Santiago, CFax: +56 2 594 2001.

E-mail address: [email protected] (C.

Acid mine drainage (AMD) from the ZnPb(AgBiCu) deposit of Cerro de Pasco (Central Peru) andwaste water from a Cu-extraction plant has been discharged since 1981 into Lake Yanamate, a naturallake with carbonate bedrock. The lake has developed a highly acidic pH of 1. Mean lake water chemistrywas characterized by 16,775 mg/L acidity as CaCO3, 4330 mg/L Fe and 29,250 mg/L SO4. Mean trace ele-ment concentrations were 86.8 mg/L Cu, 493 mg/L Zn, 2.9 mg/L Pb and 48 mg/L As, which did not differgreatly from the discharged AMD. Most elements showed increasing concentrations from the surface tothe lake bottom at a maximal depth of 41 m (e.g. from 3581 to 5433 mg/L Fe and 25,609 to 35,959 mg/LSO4). The variations in the H and O isotope compositions and the element concentrations within theupper 10 m of the water column suggest mixing with recently discharged AMD, shallow groundwaterand precipitation waters. Below 15 m a stagnant zone had developed. Gypsum (saturation index,SI 0.25) and anglesite (SI 0.1) were in equilibrium with lake water. Jarosite was oversaturated(SI 1.7) in the upper part of the water column, resulting in downward settling and re-dissolution inthe lower part of the water column (SI 0.7). Accordingly, jarosite was only found in sediments fromless than 7 m water depth. At the lake bottom, a layer of gel-like material (90 wt.% water) of pH 1 witha total organic C content of up to 4.40 wet wt.% originated from the kerosene discharge of the Cu-extrac-tion plant and had contaminant element concentrations similar to the lake water. Below the organic layerfollowed a layer of gypsum with pH 1.5, which overlaid the dissolving carbonate sediments of pH 5.37.In these two layers the contaminant elements were enriched compared to lake water in the sequenceAs < Pb Cu < Cd < Zn = Mn with increasing depth. This sequence of enrichment was explained by thefollowing processes: (i) adsorption of As on Fe-hydroxides coating plant roots at low pH (up to3326 mg/kg As), (ii) adsorption at increasing pH near the gypsum/calcite boundary (up to 1812 mg/kgPb, 2531 mg/kg Cu, and 36 mg/kg Cd), and (iii) precipitation of carbonates (up to 5177 mg/kg Zn and810 mg/kg Mn; all data corrected to a wet base). The infiltration rate was approximately equal to the dis-charge rate, thus gypsum and hydroxide precipitation had not resulted in complete clogging of the lakebedrocks.

2010 Elsevier Ltd. All rights reserved.

1. Introduction

Management of acid mine drainage (AMD) is a major problemfor the present day mining industry. Sulfide-rich deposits tend toproduce highly acidic waters, if no carbonate minerals neutralizethe effluents (Plumlee and Logsdon, 1999). Pit lakes have becomea subject of increasing importance in environmental geochemistry,

ll rights reserved.

r associates S.A., Av.11 dehile. Tel.: + 56 2 594 2024;

Wisskirchen).

due to the closure of open pit mines (Castro and Moore, 2000;Geller et al., 1998; Snchez Espaa et al., 2008). The main sourcesof acidity and metals in pit lakes include oxidation of sulfides fromcoal and metalliferous deposits or backfill, water from intercon-nected flooded underground mines or dissolution of sulfates bywater flushes (Bowell and Parshley, 2005; Castendyk et al., 2005;Denimal et al., 2005; Knller et al., 2004; Pellicori et al., 2005). Inmost pit lakes, the sources of contaminants are diffuse and difficultto apportion. In pit lakes with carbonate wall rocks or carbonate-containing veins the system is buffered to slightly basic pH,limiting the solubility of most metals (Shevenell et al., 1999).
http://dx.doi.org/10.1016/j.apgeochem.2010.04.015mailto:[email protected]://www.sciencedirect.com/science/journal/08832927http://www.elsevier.com/locate/apgeochem

1108 C. Wisskirchen et al. / Applied Geochemistry 25 (2010) 11071119

Conclusive interpretation by modeling of data sets for a number ofpit lakes has suggested control of aqueous element concentrationsby mineral equilibrium (Eary, 1999). Addition of lime is still theonly successful method to increase the alkalinity of highly acidicpit lakes. Addition of nutrients (phosphate or organic matter) tostimulate the microbiological activity in order to increase the pHhas only been applied with good results for pit lakes with pH >3(Koschorreck et al., 2007).

For AMD neutralization, carbonate minerals are an alternativeto lime, because of their lower cost. Several studies have beencarried out in recent years to investigate the interactions of car-bonates with mine effluents, and their application to AMD neutral-ization and metal retention (Komnitsas et al., 2004; Simon et al.,2005; Webb and Sasowsky, 1994). For low mineralized AMD,anoxic and oxic limestone drains have been used successfully(Cravotta and Trahan, 1999; Santomartino and Webb, 2007). How-ever, the problem of armoring of the carbonate grains and cloggingof the limestone drains by precipitation of Fe- and Al-hydroxidesand gypsum still limits the life-span of these treatment systems(e.g. Cravotta, 2008; Rose et al., 2004).

In this article, results are presented of a detailed geochemicalstudy of a natural lake within a carbonate bedrock used as anAMD storage facility for the polymetallic ZnPb(AgBiCu) de-posit Cerro de Pasco (Central Peru). The AMD with a pH 1 hasbeen discharged without treatment into Lake Yanamate (3 km2)since 1981. The formation of AMD at Cerro de Pasco and its han-dling has been studied with regard to the Excelsior waste rockdump and tailings partly underlying it (Dold et al., 2009; Smudaet al., 2007). Additionally, spoil waters with high concentrationsof kerosene from a solvent extraction-electrowinning (SX-EW)plant were discharged into the lake until August 2001. The geo-chemical approach used in this first case study of a discharge pondin carbonate rocks combines major and trace element concentra-tion of waters and sediments, H and O stable isotopes of waters,and mineralogical data. The code PHREEQC V2 (Parkhust andAppelo, 1999) was used to model element speciation and mineralprecipitation or dissolution in the water column and at the

Fig. 1. (A) Location map of the Cerro de Pasco region in Peru. (B) Map of the local geologLake Huaygacocha with sampling points. (D) Cross-section of the lake basin from west tlevel being 14 m higher than before discharge began in 1981.

watersediment interface. The new data set and geochemical mod-eling provided insights into the lake chemistry, in particular theformation of secondary minerals, the interactions in the AMD-car-bonate system, and the migration of contaminant elements.

2. Site description

Lake Yanamate is located 5 km south of the Cerro de Pasco de-posit at the northern end of the Puna plateau between the Westernand Eastern Cordillera of Peru (Fig. 1A and B). This lake is part ofthe mining facilities of the Cerro de Pasco mine, at present oper-ated by Volcan Corporation S.A.A. and previously (until 1999) bythe former state company Centromin S.A. Currently the deposit isexploited by means of both underground and open pit mining. Cer-ro de Pasco is one of the largest polymetallic resources of theworld. Accumulated past production and known reserves include175 Mt @ 7% Zn, 2% Pb and 3 oz/t Ag. Further, 1.2 billon oz ofAg and 2 million oz of Au, and an equivalent of 50 Mt of Cu havebeen produced, largely before 1950 (Baumgartner et al., 2008). Thedeposit is dominated by a 1800 300 m quartz-pyrite bodyformed by replacement of Triassic limestone, representing90 vol.% of the entire ore body (Enaudi, 1977). Different types ofpyrite have been described, some of them containing up to10 wt.% As (Petersen, 1965; Ward, 1961). The main ore mineralsare sphaleritegalena, pyrrhotite, enargite, luzonite, tennantite,marmantite, arsenopyrite and chalcopyrite (Enaudi, 1977). Finalhydrothermal and hypogene activities resulted in mineral associa-tions dominated by pyritehematiterealgar and covellitebornite,respectively. Supergene processes formed an enrichment blanketof chalcocite and covellite (Petersen, 1965). The sources of minewater were: (1) groundwater of the Cerro de Pasco fault zone infil-trating into the open pit and the underground mine and (2) rain-water accumulating in the open pit, that subsequently infiltratedinto underlying mine works (Luna Bernal and Delgado Venero,1985). Old mine works on Cu minerals were used for in situ leach-ing with highly acidic solutions. The Cu was extracted in a SX-EWplant using kerosene as a solvent with addition of 4 vol.% Acorga

y, the open pit, discharge pipe, and Lake Yanamate. (C) Map of Lake Yanamate ando east. The two lakes were connected during the sampling period due to the water

C. Wisskirchen et al. / Applied Geochemistry 25 (2010) 11071119 1109

P-5100 or M-5615, both being Cu-selective extractants (USEPA,1994). The industrial spoil solutions were pumped into the lakefrom 1981 until August 2001.

Discharge of AMD and spoil solutions resulted in an esti-mated lake volume increase from an original 3 to 25 106 m3and the inundation of a swell, which formerly separated LakeHuaygacocha from Lake Yanamate (Fig. 1C and D). The bedrockof Lake Huaygacocha consists of limestone conglomerate witha calcite matrix. The bedrock of Lake Yanamate consists of thickbedded limestone with layers of dolomite, sandy limestone and afew shale interbeds (Jenks, 1951). The two formations are sepa-rated by the Cerro de Pasco-fault. The rocks are fractured due tofaulting, and are affected by the formation of karst. Glacial activ-ity eroded loose sediments, allowing an almost direct contact ofthe lake water with the carbonates. An end moraine forms thenatural dam SW of the lake and was heightened by the construc-tion of an artificial dam, to prevent drainage to the marshlandaround Lake Cuchis. The lake covered an area of 3 km2 withina hydrologic basin of 10 km2. The south-eastern part of the lakereceives surface runoff from a marsh.

The climate of the Cerro de Pasco area is characterized by coldand dry austral winters (June to August) and warmer austral sum-mers with heavy rain events (October to March). The annual meanprecipitation of 1025 mm/a is nearly equaled by the annual meanevaporation of 988 mm/a. The annual mean temperature is4.2 C. The main wind direction is NE to SW.

3. Methods

3.1. Sampling and field methods

During sampling in June/July 2003 (dry and cold season) theAMD discharge rate was 9.5 m3/min. The lake water table was at4359 m above sea level (m a.s.l.) (14 m higher than in 1981), andhad a maximum depth of 41 m. The discharged AMD was sampledfrom a retention tank at the pumping station. From the lake, a totalnumber of 45 water samples were taken at 11 sampling stations inprofiles of 0.5 m, 5 m, and continued in 5 m steps down to the lakebottom (Fig. 1C). Samples were taken with a peristaltic pump andfiltered through 0.45 lm in-line filters, into filled 60 mL HDPE bot-tles, sub-samples for cation analysis were acidified with HNO3(Merck suprapure), and frozen immediately until analysis. Temper-ature, pH, Eh (calculated to a standard H electrode), and electricalconductivity (EC) were measured directly in the field with twoclosed flow-through cells and a set of WTW electrodes and meters.Hot acidity (H+ + hydrolysable metals) was measured within thenext 12 h by titration to pH 8.3 with 1.6 N NaOH after addition ofH2O2 and boiling, and reported as mg/L CaCO3 (Ficklin and Mosier,1999).

Lake sediments were sampled by gravity coring at the same sta-tions where water samples were taken. The 12 cores were dividedinto 44 sub-samples. Nine cores of the soils of the inundated zone(due to the increased lake water level) were taken by percussiondrilling, and were divided into 41 sub-samples. The cores were de-scribed and divided on the basis of color and grain size. Pore waterpH was measured with a Sentix SP electrode and a WTW meter.Samples were dried at 30 C and their weight was determined be-fore and after drying to calculate water content. Two samples of asoil profile on Pucar limestone and two samples of the Pucarlimestone were taken as reference for background concentrationsof contaminant elements.

To study the direct interaction of the lake water with the under-lying bedrock, fresh hand specimens of Pucar limestone taken inthe surroundings of the lake were placed in the lake at 0.5 m waterdepth for 25 days and subsequently studied.

3.2. Analytical methods

3.2.1. Isotope analysis of d2H and d18O in waterThe stable H and O isotope analyses of the water samples

were performed at the Stable Isotope Laboratory of the Univer-sity of Lausanne. The stable H isotope analyses were performedusing a ThermoFisher H-Device connected to a Delta S isotope ra-tio mass spectrometer (IRMS). For this method, the H2 gas wasproduced by reduction of a volume of 1.2 lL water over hot(840 C) Cr within a reactor connected to the dual inlet systemof the IRMS. Oxygen isotope analyses were made through equil-ibration of 0.5% CO2 in He with 1.2 mL of water for 24 h at roomtemperature followed by extraction in a continuous He flowusing a ThermoFisher GasBench II connected to a Delta Plus XLIRMS. The stable H and O isotope ratios are reported in the dnotation as the deviation relative to the Vienna StandardMean Ocean Water (VSMOW). The reproducibility, assessed interms of the within-run replicate analyses of laboratory stan-dards, was better than 0.3 and 0.1 (1r) for d2H and d18O val-ues, respectively.

3.2.2. Hydrogeochemical analysesAll the water analyses were performed at the Helmholtz Centre

for Environmental Research (UFZ) Magdeburg, Germany. Watersamples were analyzed by inductively coupled plasma-opticalemission spectrometry (ICP-OES; Perkin Elmer Optima 3000) andwith inductively coupled plasma-mass spectroscopy (ICP-MS; Agi-lent 7500 C) for major and trace elements. ICP-OES measurementsof S were used to calculate SO4 concentrations in water. DissolvedP was determined photometrically with ammonium-molybdate(German standard method DIN EN 1189). Reproducibility andaccuracy of all methods was better than 5%.

3.2.3. Chemical analyses of solid samplesThe major and trace element analyses of sediment and soil sam-

ples were performed at the Helmholtz Centre for EnvironmentalResearch (UFZ), Dept. of Analytical Chemistry in Leipzig, Germany.Aliquots of the sample material were dried at 105 C, ground withan agate ball mill, mixed with stearine wax, pressed to pellets andanalyzed for trace elements by means of the energy dispersiveX-ray fluorescence spectrometer XLAB2000 (Spectro Instruments)running the software package X-Lab Pro 2.2. The calibration func-tions for the individual analytes were based on the combination ofthe measured response and the certified concentration data ofabout 50 reference materials including soils, river- and lake sedi-ments and samples rich in Fe (Morgenstern et al., 2001). Samplesstrongly enriched with metals were diluted with SiO2 powder inorder to fit their analyte concentrations to the working ranges ofthe available calibrations. To achieve optimum spectrometer oper-ating conditions, the samples were excited by both polarized -andmonochromatic X-radiation. An aliquot of the powdered samplewas diluted with Li2B4O7 to prepare glass discs for analyzing majorelements. The measurements were performed with the wave-length dispersive X-ray fluorescence spectrometer (WDXRF) S4Pioneer (Bruker-axs). The calibrations of the WDXRF-spectrometerfor the major elements were adjusted using the certified referencematerials, CANMET-LKSD1-LKSD4 (lake sediments), CANMET-STSD1-STSD4 and GBW07309-11 (stream sediments), NIST-SRM2689 and NIST-SRM2691(coal fly ashes). The relative precisionof the measurements was better than 3% for the traces and betterthan 1% for the major components.

Total C (TC) and total organic C (TOC) were determined with aCoulomat Strhlein 702 CS in the Centre of Mineral Analysis atthe University of Lausanne, Switzerland. To determine TOC, sam-ples were moistened with ethanol and fumed-off overnight at120 C after addition of HCl. Total inorganic C (TIC) was calculated


by difference. Calibration was checked every eight sampleswith pure CaCO3, and reproducibility and accuracy was betterthan 3%.

Results of XRF and C determination were corrected for watercontent to obtain element concentrations for non-dried samples,in order to allow the direct comparison of element concentrationsin sediments and water.

3.2.4. Mineralogical analysisSolid samples were analyzed for their mineralogical composi-

tion by means of X-ray diffraction (Seifert XRD) and optical micros-copy at the Institute of Mineralogy, Crystallography and MaterialScience of the University of Leipzig. Scan settings were 0.05 2h(range 5100) and 2 s. counting time per step with monochromat-ed Cu Ka X-radiation with secondary graphite monochromator.Polished thin sections were prepared using epoxy resin to embedloose grain samples, and analyzed by optical microscopy.

3.3. Geochemical modeling

The geochemical modeling code PHREEQC V2 (Parkhust andAppelo, 1999) and the thermodynamic database minteq fromMINTEQA2 (Allison et al., 1991) were used to calculate speciationof elements in water and mineral saturation indexes. Ionicstrengths were between 0.2 and 0.7, which is in the upper rangewhere the ion-association model and Debye Hckel expressionsfor activity correction are still applicable (Parkhust and Appelo,1999).

4. Results and discussion

4.1. Isotope geochemistry of lake water

The water of the discharged AMD had a d18O value of 14.7and a d2H value of 105.7 and plots close to the Local MeteoricWater Line (LMWL) of La Paz (Fig. 2A). Spring waters feeding LakeJunin, located 30 km south of Cerro de Pasco at 4100 m a.s.l., hadcomparable values with d18O from 15.8 to 13.5 and d2Hfrom 116.0 to 102.1 (Flusche et al., 2005). This indicates alocal origin of the discharged mine water.

The d18O values of the lake waters ranged from 11.7 to8.7 and the d2H values ranged from 93.9 to 79.7. Alllake water samples were isotopically heavier than the dischargedAMD, and followed an evaporation line (d2H = 4.5 d18O 40). Evap-oration seems to be the main process controlling the variation ofthe d2H and d18O values in the lake. The slope of 4.5 indicates a

Fig. 2. (A) d2H v/s d18O binary plot of lake water samples, discharged AMD, and the Locselected water columns of the lake.

humidity of 60% (Gonfiatini, 1986), which is lower than the aver-age value of 74% reported in the literature for Cerro de Pasco(PAMA, 1996). The AMD plots between the evaporation line andthe LMWL of La Paz. The AMD waters originated mainly fromfault-waters of the Cerro de Pasco fault zone and precipitationinfiltrating the underground mine.

From 15 m water depth to the surface, the mean d18O value in-creased from 10.6 to 9.6 and the mean d2H value from88.7 to 82.5 (Fig. 2B). The increases in d2H and d18O valuesfrom 15 m to 10 m and from 5 m to 0.5 m were nearly equal (0.17d18O/m and 0.21 d18O/m, respectively). The uppermost waters(0.5 and 5 m depth) had a mean d18O value shifted 1.6 to a morenegative value. This isotope shift can be explained by mixture ofthe lake upper 5 m with recent AMD discharge and precipitation.The deep waters of the lake (2035 m) had a relatively uniform iso-tope composition with d18O values 11.1 and d2H values93.6 (Fig. 2B), indicating a stagnant bottom stratum (termedmonimolimnion), which was excluded from taking part in lakewater mixing and circulation processes.

The isotopically heaviest water (d18O 8.7 and d2H 79.7)was at CPY14(10 m). This isotopically heavier water suggested a lo-cal contribution from the marshes SE of the lake (Fig. 1B).

4.2. Hydrogeochemistry

The AMD discharged into the lake had pH 1.1, Eh 609 mV, andan EC of 27.4 mS/cm (Table 1). The concentrations of Fe and SO4were 4440 mg/L and 31,875 mg/L respectively, and acidity was18,720 mg/L as CaCO3. Copper had a concentration of 125 mg/L,Zn 748 mg/L, Pb 4.2 mg/L and As 47.8 mg/L. These values wereclose to the range reported for AMD from massive pyrite deposits(Plumlee, 1999) and fall into the class of extremely acidic and ex-tremely mineralized pit lakes, following the classification of Sn-chez Espaa et al. (2008).

The average hydrogeochemical parameters determinedfrom the 45 lake water samples were pH 1.0, Eh 653 mV, EC24.1 mS/cm, 4330 mg/L Fe, 29,250 mg/L SO4, 16,775 mg/L acidityas CaCO3, 87 mg/L Cu, 493 mg/L Zn, 3 mg/L Pb and 48 mg/L As.These mean values did not differ greatly from values measuredfor the discharged AMD.

The pH and Eh decreased slightly with depth from pH 1.2 and672 mV at 0.5 m to pH 1.1 and Eh values between 585 and652 mV at the lake bottom (Fig. 3A). Acidity increased from15,712 mg/L to 19,175 mg/L as CaCO3 with depth, indicating thatdissolution of carbonate bedrock did not result in higher pH andlower acidity at the lake bottom. Compared to the value of27.4 mS/cm for discharged AMD, the upper 15 m of the lake water

al Meteoric Water Line of La Paz (LMWL La Paz). (B) d18O and d2H in VSMOW of

Table 1Stable isotope data of d2H and d18O, physicochemical parameters, and concentrations of selected elements in the sampled water columns of Lakes Yanamate and Huaygacocha and the Cerro de Pasco AMD. The calculated meanconcentration and standard deviation includes all water samples of both lakes. Abbreviations: EC = electrical conductivity; acy = acidity as CaCO3; n.a. = not analyzed; Stdev = standard deviation.

Column Depth(m)

d18O()

d2H()

pH()

Eh(mV)

EC(mS/cm)

acy(mg/L)

Na(mg/L)

K(mg/L)

Mg(mg/L)

Ca(mg/L)

Mn(mg/L)

Al(mg/L)

Fe(mg/L)

SO4(mg/L)

Cu(mg/L)

Zn(mg/L)

Pb(mg/L)

As(mg/L)

Si(mg/L)

P(mg/L)

Ti(lg/L)

V(lg/L)

Cr(lg/L)

Co(lg/L)

Ni(lg/L)

Zr(lg/L)

Cd(lg/L)

Bi(lg/L)

U(lg/L)

CPY04 0.5 9.4 82.0 1.0 673 19.5 n.a. 16.2 30.9 221 383 178 248 3920 29,865 91.1 477 2.7 48.3 14.0 33 515 1936 181 156 299 66 1235 1708 1105 10.2 83.3 1.0 661 21.7 n.a. 16.0 29.8 208 390 170 244 3950 29,340 86.1 467 2.8 47.2 13.0 36 480 1924 182 166 297 54 1218 1498 10710 9.6 82.3 1.0 650 23.4 15,840 17.1 31.1 216 419 176 265 4430 31,380 82.9 493 2.5 50.7 14.5 36 431 2082 199 189 325 44 1347 1030 12014.5 10.1 85.8 1.0 621 25.9 n.a. 17.4 31.2 214 448 172 279 4990 33,300 68.7 487 2.2 56.2 15.3 39 346 2188 214 221 348 38 1426 540 117

CPY05 0.5 n.a. n.a. 1.3 674 20.0 n.a. 16.5 31.1 219 378 179 248 3950 28,935 90.9 482 3.0 52.1 13.9 35 563 2149 204 171 318 49 1319 1916 1212.5 n.a. n.a. 1.5 673 19.9 n.a. 16.5 30.9 218 377 179 247 3940 28,485 90.3 479 2.8 47.9 13.6 35 523 1926 178 159 289 96 1239 1742 112

CPY07 0.5 n.a. n.a. 1.0 673 18.9 n.a. 16.6 30.9 219 378 179 248 3950 28,470 90.7 482 2.7 47.6 13.9 34 517 1918 180 161 281 58 1232 1763 1156 n.a. n.a. 0.9 654 22.2 n.a. 16.2 29.4 201 381 164 244 4050 28,005 78.8 n.a. 2.6 50.1 12.1 36 439 2068 194 182 301 39 1336 1097 120

CPY09 0.5 n.a. n.a. 1.5 673 18.1 15,962 17.0 31.7 224 384 184 251 4020 29,025 92.8 493 2.9 48.1 16.4 36 523 1952 183 160 290 48 1276 1803 1165 n.a. n.a. 1.1 660 21.1 16,191 17.1 31.8 224 383 185 252 4050 29,325 96.8 500 3.5 48.8 14.8 36 515 1981 187 164 302 44 1266 1925 11710 n.a. n.a. 1.2 657 23.2 17,228 18.4 34.5 233 397 196 278 4540 32,340 100.0 534 3.0 54.3 17.0 38 541 2121 192 180 316 41 1362 1930 12815 n.a. n.a. 1.1 645 27.4 19,302 19.5 35.9 239 421 202 304 5330 36,150 96.4 557 2.8 59.3 18.4 39 502 2296 218 208 341 42 1474 1475 13320 n.a. n.a. 1.1 637 29.4 19,860 19.5 35.4 227 413 191 300 5480 37,050 86.8 542 2.8 61.1 19.4 40 420 2333 228 224 354 35 1509 1110 13625 n.a. n.a. 1.1 637 29.5 19,701 19.6 35.4 228 399 192 303 5520 37,050 86.8 549 2.8 62.4 18.3 39 454 2390 224 237 355 35 1516 1123 140

CPY10 0.5 n.a. n.a. 1.0 673 21.0 14,608 14.5 28.0 201 367 162 221 3550 25,515 81.9 439 2.7 36.8 13.3 32 484 1772 165 144 266 82 1150 1745 1055 n.a. n.a. 0.8 659 23.8 17,600 18.0 32.8 235 389 195 266 4240 29,985 99.2 517 3.3 51.1 32.6 37 554 2050 192 168 305 62 1336 1908 12210 n.a. n.a. 0.8 657 25.9 16,320 18.4 34.6 231 396 194 276 4520 32,025 99.6 527 3.0 53.9 15.0 39 562 2114 195 178 308 48 1376 1912 12815 n.a. n.a. 0.7 651 27.8 16,400 18.6 35.0 235 393 197 282 4720 33,075 102.0 532 3.0 57.4 16.2 40 584 2251 213 189 331 47 1439 1992 13118 n.a. n.a. 0.6 637 31.3 21,440 19.2 35.0 223 405 188 293 5330 35,700 85.7 537 2.8 59.2 19.1 42 429 2276 218 217 343 35 1458 1119 131

CPY13 0.5 9.3 82.3 1.6 673 18.5 14,800 17.2 31.9 225 385 184 252 4020 28,440 92.8 499 2.8 47.0 14.1 37 521 1922 181 157 289 36 1239 1794 1165 10.0 85.0 1.2 660 21.1 15,120 15.4 28.2 203 344 167 226 3630 25,635 86.2 453 2.6 33.1 12.3 30 472 1747 159 140 270 29 1140 1715 10410 10.2 85.8 1.2 657 24.6 16,480 18.9 35.2 238 406 200 284 4640 32,190 102.0 547 3.0 54.1 15.4 40 562 2147 201 183 311 40 1386 1930 12615 10.3 86.7 1.2 652 25.1 17,040 18.7 35.4 236 397 200 285 4760 33,135 103.0 546 5.3 56.7 16.2 41 539 2215 207 185 321 41 1398 1978 129

CPY14 0.5 9.8 83.6 1.1 673 19.5 15,696 17.1 31.4 226 382 182 250 3990 28,110 92.1 494 2.8 48.1 16.4 36 510 1940 176 158 284 34 1266 1801 1155 11.3 86.6 1.5 659 22.3 16,928 14.9 26.9 192 383 159 216 3520 25,005 83.0 432 2.7 31.6 13.9 30 443 1649 152 134 248 23 1079 1740 9510 8.6 79.7 1.4 657 24.2 17,248 9.7 16.5 114 362 95 130 2270 16,365 49.9 268 2.8 19.8 7.1 18 260 1022 94 91 160 7 676 1439 6015 9.8 86.7 1.4 654 25.4 18,272 18.6 35.1 239 394 198 281 4680 32,880 103.0 540 2.9 55.8 17.4 33 565 2176 207 186 326 39 1383 1960 126

CPY16 0.5 10.2 83.2 1.4 672 19.1 15,920 17.4 32.0 226 388 186 254 4050 28,620 93.6 502 2.7 47.7 15.1 34 510 1960 183 159 292 30 1255 1801 1125 10.9 86.2 1.3 660 22.9 16,160 15.2 27.3 195 391 161 221 3550 14,085 83.7 414 2.7 32.0 15.4 30 440 1668 158 132 249 71 1057 1730 9610 10.7 85.0 1.2 656 24.9 17,792 18.6 35.1 234 401 197 280 4600 14,205 101.0 536 3.0 54.4 14.7 38 539 2121 195 180 326 57 1371 1871 12517 11.7 91.7 n.a. 636 31.3 19,296 19.1 35.1 222 405 187 292 5330 35,550 85.6 532 2.7 61.1 18.8 40 446 2327 227 225 357 39 1462 1132 134

CPY19 0.5 9.8 83.3 0.7 673 18.9 14,480 17.1 31.7 223 384 183 251 4000 28,350 92.6 497 2.8 47.1 13.9 34 507 1904 177 155 282 32 1219 1737 1105 10.5 87.2 0.6 660 22.1 14,880 17.7 32.6 232 387 192 263 4190 29,490 99.3 511 2.9 50.6 14.1 35 536 1997 189 163 298 35 1297 1884 11810 9.8 84.5 0.6 657 24.6 16,512 18.5 34.7 232 398 195 277 4530 31,650 100.0 532 2.8 54.4 15.2 38 554 2159 199 181 318 41 1369 1906 12516 11.2 91.5 0.5 641 28.8 18,320 18.9 35.0 228 412 189 294 5230 35,100 87.6 540 2.8 60.2 17.7 39 482 2314 215 219 343 34 1477 1225 133

CPY20 0.5 8.9 80.9 1.1 672 19.3 15,360 7.4 11.1 80.6 392 64 84 1440 11,205 33.5 180 2.8 12.7 5.9 12 170 686 63 59 105 n.a. 450 1434 385 10.3 86.1 1.1 662 21.0 15,840 11.9 20.2 148 376 121 163 2670 19,050 62.7 332 2.8 24.4 12.3 22 339 1286 120 107 195 11 824 1637 7310 9.6 85.5 1.0 657 23.9 16,800 19.3 36.4 243 419 205 291 4760 33,255 105.0 564 2.8 54.2 20.5 37 524 2104 200 178 310 33 1346 1834 12315 10.6 89.7 1.0 650 26.1 17,920 17.7 33.2 227 401 191 267 4710 32,355 99.6 529 2.9 53.8 17.3 37 495 2074 197 195 322 31 1331 1747 11820 11.2 93.9 0.9 637 30.1 18,800 16.3 29.1 204 413 171 242 5030 32,340 82.6 503 2.6 49.0 17.3 32 340 1871 178 214 320 19 1256 986 10625 11.2 93.9 1.0 631 30.6 20,240 16.6 27.4 185 403 156 229 4760 30,375 74.0 473 2.6 46.6 19.4 31 329 1833 176 199 302 14 1219 956 9930 10.9 92.8 1.0 589 30.4 20,992 40.1 35.4 208 459 188 371 7010 41,250 76.1 728 2.8 63.1 22.9 44 412 3251 367 342 555 20 2357 800 13335 11.1 93.6 1.0 588 29.5 20,160 40.8 36.4 214 466 193 384 7180 42,300 78.1 746 2.6 58.1 21.7 33 415 3018 338 323 524 15 2240 756 126

CPY27 0.5 n.a. n.a. 1.8 673 18.9 16,000 11.9 20.5 148 338 120 164 2660 18,945 60.3 328 2.5 23.8 14.1 22 330 1265 122 102 188 8 798 1392 733.5 n.a. n.a. 1.4 673 18.6 15,840 13.2 23.5 167 327 135 183 3010 21,345 69.3 375 2.4 27.4 11.2 26 381 1444 140 121 207 15 931 1444 83

MEANlake

10.3 86.4 1.1 654 23.8 17,193 17.8 30.9 211 394 176 255 4327 29,250 86.8 493 2.8 48.0 15.8 34 467 1996 190 177 304 39 1296 1555 113

Stdevlake

0.8 4.1 0.3 20 4.1 1865 5.6 5.4 33 26 29 53 1042 6749 15 95 0.4 12.0 4.1 6 89 427 49 50 73 18 310 388 21

AMD 14.7 105.7 1.1 609 27.4 18,720 19.1 40.6 375 391 287 284 4440 31,875 125.0 784 4.2 47.8 24.2 0.1 660 2304 209 156 348 44 1615 3428 172

C.Wisskirchen

etal./A

ppliedG

eochemistry

25(2010)

11071119

1111

Fig. 3. (A) pH, Eh, electrical conductivity (EC), acidity (Acy). (B,C,D) Element concentration distribution of Na, K, Mg, Ca, Mn, Al, Fe, SO4, Cu, Zn, Pb and As in 11 water columnssampled from Lake Yanamate and Lake Huaygacocha. (E) Saturation indices (SIs) for gypsum, anglesite, jarosite and strengite, calculated with PHREEQCI V2 and the minteqdatabase. Vertical dashed lines mark values for inflowing AMD.



column had mean values of 19 mS/cm, while the bottom watershad equal or higher values. Below 20 m water depth all the EC val-ues were around 30 mS/cm.

From the top to the bottom of the water body, mean concentra-tions of Fe increased by 52% from 3581 mg/L to 5433 mg/L and forSO4 by 40% from 25,609 mg/L to 35,959 mg/L (Fig. 3C). Aluminumfollowed a trend similar to that of Fe. Calcium concentrations in thelake (range 327466 mg/L; mean 397 mg/L) were equal to dis-charged AMD (391 mg/L). In contrast, Mg was depleted in the lake(range 81243 mg/L, mean 214 mg/L) compared to dischargedAMD (375 mg/L). Gypsum (CaSO42H2O) had relatively invariablesaturation indices (SI) around 0.25 (Fig. 3E) and controlled theconcentration of Ca in the lake and the AMD. Jarosite(KFe33 OH6SO42) was oversaturated (SI = 0.8 to +4.5, mean1.7) in the upper samples and undersaturated in the deepest sam-ples of the water body (SI = 3.8 to +2.3, mean 0.7) (Fig. 3E).Potassium was depleted in all lake water samples (range 1136 mg/L, mean 31 mg/L) compared to 40.6 mg/L K for dischargedAMD, while Na concentrations of the lake waters were close tothe 19 mg/L measured for discharged AMD for most samples, andonly depleted for some samples near the lake surface (Fig. 3B). Thiscan be explained by partial removal of K from the water column byjarosite precipitation. Jarosite was found by XRD only in sedimentsof water depth 67 m; this supports the hypothesis of element cy-cling by jarosite (see Section 4.5.). Its formation in the upper part ofthe water column and its downward settling and later dissolutionmay have resulted in downward element cycling of Fe and SO4.Redox cycling of Fe in natural lakes is a well-studied subject(Hamilton-Taylor and Davision, 1995; Hamilton-Taylor et al.,2005). The formation of schwertmannite (2.75 mol H+ per molFe3+) or jarosite (2 mol H+ per mol Fe3+) liberates acidity, thus theseminerals can control the lake pH, as observed in the Berkeley pitlake and several pit lakes of the Iberian Pyrite Belt for schwertman-nite (Pellicori et al., 2005; Snchez Espaa et al., 2008), e.g.:

3Fe3 K 2SO24 6H2O() KFe33 OH6SO42 6H

1

In Lake Yanamate, however, the assumed dissolution of jarositein the lower part of the water had no apparent effect on the watercomposition. The acidity consumption was probably obscured by ahigher acidity and mineralization in the lower part of the lakewater, due to its origin from times before the discharge of spoilsolutions from the SX-EW plant was stopped.

The discharged AMD contained almost no dissolved P (0.1 mg/L).Concentrations of 35 mg/L dissolved P in the lake water originatedfrom the dung of llamas and alpacas feeding frequently close to thelake, and entered the lake with the surface runoff (Table 1). Streng-ite (FePO42H2O) was found to be close to saturation (SI = 0.91.1(mean 0.1)) in the upper part of the water column and undersatu-rated in the lower part of the water column (SI = 1.60.4(mean 0.5)) and may have played a smaller role in elementcycling.

Manganese seemed to have precipitated partially from thewater column as its concentration is lower in the lake (range 64205 mg/L; mean 176 mg/L) than the 287 mg/L measured in AMD.Copper showed increasing concentrations from 85.2 mg/L at thetop to 93.0 mg/L at 10 m water depth and decreased below to amean value of 87 mg/L for the bottom layer, possibly due to thepresence of the Cu-selective extractant in the mixed dischargeand preferred settlement (Fig. 3D). Zinc and Cd concentrations in-creased slightly with depth (from 455 to 560 mg/L Zn and from 1.2to 1.5 mg/L Cd). Lead had a uniform concentration of 2.8 mg/L(Fig. 3D). Anglesite (PbSO4) had a SI of 0.08 and controlled the con-centration of Pb in the lake water, causing its spatially invariableconcentration (Fig. 3E). The element of greatest environmentalconcern in the lake was As with a mean of 48 mg/L, showingincreasing mean concentrations with depth from 41 to 61 mg/L.

The Eh decreased stronger in Lake Huaygacocha from 673 to621 mV than in Lake Yanamate. Vertical mixing of lakes is partlycontrolled by their relative depths (maximum depth divided bywidth) (Castro and Moore, 2000), which is higher for Lake Huayg-acocha and can explain the differences between the lakes hydrog-eochemistry. Concentrations of metals were slightly lower than inLake Yanamate (CPY04 and CPY07, Fig. 3D).

In the austral winter of 2003 Lake Yanamate was a meromicticlake without complete mixing and its density gradient was domi-nated by the high mineralization of the water. The monimolimnionfound for CPY09 and CPY20 below 20 m probably represented themixed discharge of AMD and the spoil water of the SX-EW plant.Calculated with the discharge rate of July 2003, the lake waterbody has not been completely replaced in the time after the SX-EW plant was closed in 2001. Possibly in the austral summer, thehigher precipitation may result in an epilimnion and less verticalmixing. The low element concentrations for CPY14(10 m) are inagreement with its heavier isotope signature, interpreted as a localcontribution of near-surface groundwater discharge from themarshes SE of the lake, resulting in dilution (Fig. 2B and Table 1).

4.3. Characterization of lake sediments

At the end of the discharge pipe, the discharged AMD formed a2 m deep channel in the limestone, showing the general reactivityof the bedrock with the acidic water. Parts of this area were cov-ered by tailings. Close to the discharge point (CPY11 and CPY13;Fig. 1C) pyrite, sphalerite, anglesite, quartz and muscovite settledout of the mine water in the form of detritus-rich sediment, takenup from the tailings and transported into the lake (Fig. 4A). In thesesediments, the concentrations of Fe, Cu, Zn, As, Pb and Cd were thehighest found in the lake sediments (Table 2).

In most sediment cores a sequence of gel-like matter abovefine-grained material was found. The layer of gel-like matter wasof varying thickness, between 3 and 32 cm. The color of the gelsvaried between gray and brown in Lake Yanamate, and in LakeHuaygacocha a layer of greenish-gray color was found (CPY23,CPY24). The mean water content was 89 wt.% in the liquid and79 wt.% in the rigid gels. The gels contained 1.64.4 (mean 3.0)wt.% TOC and partly imbedded residues of plants, algae and seedvessels. Some gels contained calcite and had TIC contents of upto 1 wt.% (Fig. 4B). The pH of the gels was 11.5 where no calciteresidues were found and 5.66.4 where calcite was detected(CPY14, Fig. 4B). Gypsum detected by XRD in the dried gels crystal-lized largely during the drying process of the samples and waspresent in aggregates of up to 8 mm in size, which were not foundduring sampling.

With the spoil water of the SX-EW plant, kerosene entered thelake until August 2001 (Cerro de Pasco staff, pers. comm.). Oildroplets formed through wave action or during the industrial pro-cessing can flocculate as oil-mineral aggregates (OMA), and settledown due to an increased bulk density (Lee, 2002; Muschenheimand Lee, 2002). Visibility in the lake water was around 15 cmand filter residues were high, thus suspended matter was abun-dant. The high mineralization of lake water favored the stabiliza-tion of the OMA through coagulation with cations (Bragg andYang, 1995). The density of the OMA may have allowed its wind-driven resuspension from the flooded soils and its redeposition inthe deeper zones of the lake. Aging of low-crystalline gypsum mayexplain the presence of the rigid gels below the liquid gels. The gellayers found accumulated during the 20 a of kerosene discharge.Microbial decomposition of organic matter is very slow or absentin low-pH environments (Pelmont, 2005).

Materials of clayish to sandy grain sizes with or without gel coveron top were gray to brown depending on the organic matter content,ocher and red in Fe-rich horizons, and gray in the carbonate-rich

Fig. 4. Sedimentology, mineralogy, pH, TIC, TOC, water content, and bulk chemical data (corrected to a wet base) of selected sediment cores of Lake Yanamate. For bettervisibility, graphs of trace elements have varying scales.


deeper parts of the profiles. The mean water content was lower andvaried much more than in the gels, from 14 to 84 (mean 51) wt.%.TOC was between 1.4 and 3.3 (mean 1.7) wt.% and TIC was onlyfound in samples in which dolomite or calcite was detected. Theselayers appear to originate from a former minor loose sediment layeron top of the bedrock, because reference samples of carbonate bed-rock did not contain any TOC. The pH was between 1.4 and 4.5 whereno carbonate was detected. Gypsum was absent in the detrital-en-riched layer at CPY13, and was formed by the infiltrating acidic sul-fate solutions, with dissolved calcite as a source of Ca. Wherecarbonates were present, pH increased and stabilized at around 55.6, consistent with the modeled equilibrium pH of 5.3 for averagelake water with calcite (Figs. 4 and 5).

Down to the maximum sampled depth of 69 cm, no unalteredlimestone bedrock was reached. The experiment with limestonespecimens put in the lake for about one month showed that, as a

first step, the acidic water of the lake infiltrated the fractures ofthe limestone bedrock. In thin sections, hydroxides or hydroxysul-fates were visible in the fractures. Additional to Fe- and possiblyAl-hydroxides, the formation of gypsum (replacing calcite) mayclog fractures, because gypsum occupies twice the volume of cal-cite and could decrease the infiltration rate.

4.4. Characterization of soil sediments

All soils except CPY18 had a root layer on top and only CPY07and CPY24 (both located in Lake Huaygacocha) were covered bya 10 or 40 cm thick layer of liquid gel, respectively. Soils weremainly of argillaceous grain size. Root-containing layers exhibiteda middle brown color; deeper horizons were light brown, or insome samples ochre or oxide red where high concentrations ofFe were detected.

Table 2Element concentrations of some selected sediment and soil profiles. Element concentrations were corrected to a wet base. Abbreviations: TIC = total inorganic C; TOC: total organic C; ref soil = reference soil; n.a. = not analyzed;bdl = below detection limit.

Depth(cm)

TIC(wt.%)

TOC(wt.%)

H2O(wt.%)

Al2O3(wt.%)

SiO2(wt.%)

K2O(wt.%)

CaO(wt.%)

Fe2O3(wt.%)

S(wt.%)

P2O5(mg/kg)

V(mg/kg)

Cr(mg/kg)

Mn(mg/kg)

Ni(mg/kg)

Cu(mg/kg)

Zn(mg/kg)

As(mg/kg)

Sr(mg/kg)

Zr(mg/kg)

Pb(mg/kg)

U(mg/kg)

Bi(mg/kg)

Cd(mg/kg)

CPY06 06 0.04 8.12 68 0.47 3.28 0.10 2.62 2.11 1.65 1509 5 8 120 4 464 387 1436 58 13 100 3.5 46.7 2.0614 0.49 2.36 40 0.61 2.50 0.11 15.65 2.84 6.45 355 10 bdl 111 bdl 463 368 989 210 4 87 bdl 21.0 7.61420 0.08 3.61 46 0.41 2.33 0.07 8.07 9.00 4.67 1346 25 8 103 4 346 332 2157 122 6 51 bdl 8.1 2.92026 0.31 8.85 58 1.62 4.85 0.10 4.00 3.34 2.07 1488 21 8 171 3 1391 846 265 77 16 202 4.2 bdl 21.82934 5.99 1.65 43 0.10 0.70 bdl 27.54 1.28 0.18 266 bdl bdl 539 bdl 15 2258 13 382 3 6 bdl bdl 3.0

CPY07 05 n.a. n.a. 92 0.03 0.20 0.03 0.25 0.95 0.50 94 bdl 2 80 1 47 246 81 22 3 39 bdl 63.0 0.7510 n.a. n.a. 93 0.03 0.14 0.03 0.17 0.87 0.39 72 bdl 2 57 1 36 177 79 20 3 45 bdl 60.1 0.51015 0.04 1.88 84 0.05 0.59 0.08 2.73 1.45 1.65 218 2 2 70 1 54 188 567 43 1 84 bdl 41.3 0.81518 0.34 2.14 51 0.09 0.70 0.14 11.03 4.04 5.69 699 8 bdl 59 2 127 172 2402 104 1 212 bdl 10.3 1.01825 0.70 3.18 73 0.33 1.25 bdl 7.20 0.59 1.97 309 4 5 79 1 648 888 135 88 1 128 1.2 6.2 18.93035 3.37 2.37 62 0.02 0.48 bdl 17.37 1.26 0.19 263 bdl bdl 370 bdl 17 890 13 309 bdl 3 bdl bdl bdl4550 2.74 1.50 75 0.02 0.15 bdl 11.90 0.18 0.08 83 bdl bdl 74 bdl 9 37 13 165 1 3 bdl bdl bdl

CPY10 010 0.05 1.90 92 0.07 0.36 0.01 0.44 0.71 0.68 50 1 3 137 1 132 550 62 7 1 8 0.4 1.9 2.41020 n.a. 3.64 91 0.07 0.66 0.02 0.15 0.43 0.37 81 1 2 70 1 69 314 40 4 2 3 0.7 5.9 0.92030 n.a. n.a. 88 0.12 0.93 0.03 0.11 0.74 0.50 101 2 3 135 2 60 510 64 5 3 5 0.3 51.2 1.43032 n.a. n.a. 54 0.80 4.72 0.14 9.75 0.77 4.95 990 13 10 156 2 391 575 68 112 6 16 bdl 9.7 2.33235 1.79 1.35 16 6.26 35.02 1.08 11.92 2.61 3.48 11,617 56 45 544 16 130 3084 263 274 36 22 1.8 1.7 3.63740 n.a. n.a. 84 0.32 0.87 0.03 0.72 1.04 0.69 650 3 3 154 4 19 1501 55 17 3 4 0.7 0.5 0.84045 2.12 3.25 72 0.28 1.99 0.07 9.27 1.26 0.31 372 5 3 266 3 13 1235 21 125 4 3 0.6 bdl bdl

CPY11 grabber n.a. n.a. n.a. 8.58 47.93 1.43 0.38 18.91 10.68 3800 66 41 937 13 1650 59,220 1630 730 112 11,700 2.2 474.0 231.1

CPY13 010 0.27 1.64 63 2.81 18.87 0.52 0.10 4.96 2.85 1032 19 25 138 7 814 10,323 403 303 33 3076 2.0 157.3 38.01220 0.11 3.39 76 0.21 3.16 0.04 4.23 0.57 2.33 113 6 6 77 bdl 948 476 67 69 6 142 2.1 87.8 2.32030 0.02 2.80 80 0.14 0.98 0.01 4.00 0.42 2.20 88 4 4 64 1 402 983 371 55 1 58 0.9 2.0 13.83040 n.a. n.a. 81 0.06 0.59 0.01 5.08 0.61 0.37 240 bdl bdl 84 1 20 598 27 82 1 34 bdl 2.3 0.6

CPY14 07 0.09 2.52 91 0.09 0.35 0.01 0.22 1.03 0.75 22 2 4 179 2 151 720 80 6 1 13 0.3 1.8 2.3711 0.23 4.40 89 0.05 0.28 0.02 0.21 0.44 0.34 55 4 3 67 1 231 298 53 5 1 8 1.5 3.3 0.81116 0.11 3.83 73 0.19 1.10 0.03 5.53 0.57 2.89 bdl 5 5 88 bdl 405 296 66 59 2 10 1.2 1.4 1.41618 0.06 3.54 67 1.65 1.67 0.04 6.43 0.76 3.52 3175 10 9 116 3 638 427 1908 76 3 77 1.1 1.3 2.31825 0.95 3.36 79 0.46 3.22 0.13 4.69 1.04 0.50 561 5 5 131 5 20 662 81 80 6 4 0.9 bdl 7.62530 0.81 3.69 75 1.01 6.36 0.25 4.04 1.21 0.55 927 11 9 120 9 13 598 80 96 11 6 1.1 bdl 0.5

CPY20 07 0.06 1.77 88 0.18 1.46 0.02 1.97 0.62 0.97 530 3 7 107 2 186 964 102 34 1 157 0.3 10.0 4.3714 1.22 2.14 86 0.07 1.20 0.01 4.40 0.24 0.14 411 2 bdl 54 1 19 54 10 75 1 35 bdl 1.0 bdl1421 n.a. n.a. 88 0.06 1.49 0.01 2.91 0.09 0.08 319 bdl 1 36 1 6 33 5 54 1 4 bdl 0.2 0.3

CPY23 05 0.00 4.10 85 0.11 0.47 0.01 0.83 1.21 1.21 119 2 5 283 4 252 887 109 13 4 14 0.6 56.7 2.8510 0.01 2.41 86 0.08 0.47 0.01 2.15 0.69 1.48 bdl bdl 2 153 2 127 431 64 27 2 40 0.3 41.6 1.81015 n.a. 1.41 75 0.06 0.67 0.11 4.36 3.23 2.52 1166 6 bdl 86 4 68 252 3328 79 1 181 bdl 34.6 0.91520 0.05 1.11 70 0.14 0.90 0.06 5.28 4.37 2.97 1322 16 7 104 2 144 314 1830 89 bdl 14 bdl 2.1 1.42025 0.00 1.54 61 2.62 1.96 bdl 8.05 1.68 4.36 254 23 10 121 bdl 1728 443 98 115 bdl 5 5.6 bdl 6.62530 0.12 1.54 59 4.84 2.41 bdl 6.98 2.03 4.03 bdl 25 7 437 bdl 2531 3847 18 104 bdl 8 9.8 bdl 7.2

CPY25 05 2.65 0.02 61 0.10 0.33 0.03 9.77 0.95 4.96 272 bdl 5 78 bdl 96 197 431 104 2 142 bdl 77.4 0.8510 2.72 0.86 48 0.12 0.35 bdl 14.61 0.74 6.66 303 bdl bdl 88 bdl 155 362 289 145 2 66 bdl 65.9 2.61015 4.39 3.70 49 0.08 0.44 bdl 21.63 1.05 1.10 742 bdl bdl 192 bdl 83 1324 38 349 1 111 bdl 6.4 1.51520 6.84 1.25 44 0.07 0.42 bdl 27.37 0.88 0.30 723 bdl bdl 294 bdl 39 1436 19 463 1 61 bdl 2.6 0.92027 7.45 0.07 43 0.07 0.46 bdl 27.81 1.55 0.23 642 bdl bdl 516 bdl 18 3422 29 503 2 14 bdl 3.0 6.12735 7.78 0.00 40 0.10 0.70 bdl 28.21 2.26 0.26 512 bdl bdl 808 bdl 5 5177 3 518 4 6 1.8 bdl 1.8

Ref.soil

1020 6.18 1.12 n.a. 4.50 69.86 1.13 1.46 1.37 0.11 4167 39 37 817 11 67 245 76 105 74 165 2.2 2.0 bdl

2030 3.63 1.15 n.a. 6.69 70.80 1.86 1.67 2.11 0.12 5502 65 43 1173 14 22 303 81 124 93 65 bdl bdl bdl

C.Wisskirchen

etal./A

ppliedG

eochemistry

25(2010)

11071119

1115

Fig. 5. Sedimentology, mineralogy, pH, TIC, TOC, water content, and bulk chemical data (corrected to a wet base) of selected soil cores of Lake Yanamate (CPY25 and CPY06)and of Lake Huaygacocha (CPY07), and one sediment core of Lake Huaygacocha (CPY23). For better visibility, graphs of trace elements have varying scales.



CPY06 had comparatively high TOC with 8.1 and 8.9 wt.% in theupper root-containing horizon and between 21 and 26 cm, wherethe brownish color indicated a former top-soil (Fig. 5B). CPY06 islocated in the Pampa area between the two lakes and may havehad dense vegetation before the discharge period. Other soil sam-ples had TOC contents 64 wt.%. Soil layers that were free of calcite(XRD and TIC results) had a pH of 12. In deeper layers, wherecalcite was detected as a residual (CPY06) or dominating mineral(CPY25), the pH was >5 (Fig. 5).

4.5. Jarosite in solid samples down to 7 m water depth

Jarosite was found in all the 3 soil samples from the smallerLake Huaygacocha: in liquid gel of two cores at 7 m water depth(CPY24 and CPY07) and once in the clay size matrix for a core at3.5 m water depth (CPY05). In contrast, two cores taken at 14 mwater depth near the center of Lake Huaygacocha did not showjarosite (CPY04 and CPY23). The cores CPY07 and CPY23 areshown in Fig. 5. The occurrence of jarosite as detected by XRDis in accordance with the results of modeling, which showed thatjarosite is stable in the upper part but not in the deeper part ofthe water body. It could not be explained by the modeling, whyjarosite was found only in the sediments of the smaller LakeHuaygacocha. Similar to a Na-jarosite formation in a wasteH2SO4 lake with a pH 0.85 (Schuiling and van Gaans, 1997),concentrations of contaminant metals and As contained in LakeHuaygacochas K-jarosite bearing layers were 1001000 timeslower than in jarosites from Zn industries, because the sorptionof metals is positively correlated with pH (Bigham and Nord-strom, 2000).

4.6. Trace metal and arsenic distribution in gels

Samples CPY10(010 cm) and CPY14(07 cm) (Fig. 4) had com-paratively low concentrations of Al2O3 and SiO2 in their liquid gelmatrices (0.07 and 0.09 wt.% Al2O3; 0.36 and 0.35 wt.% SiO2,respectively). Thus detrital input was minimal for these two sam-ples and geochemical data was interpreted as characteristic for thegel-like matter. Both had similar pH of 1.2 and 1.5, similar watercontent of 91 and 92 wt.%, and similar TOC of 1.9 and 2.2 wt.%,respectively. The absence of detritus allowed the element concen-trations in the liquid gels to be compared with those of the overly-ing water. CPY10(010 cm) and CPY14(07 cm) had 132 and151 mg/kg Cu, 550 and 720 mg/kg Zn, and 62 and 80 mg/kg As,respectively, thus they were not significantly enriched in these ele-ments compared to the overlying water with 85.7 mg/L Cu,537 mg/L Zn, and 59 mg/L As for CPY10(18 m) bottom water. Leadwas more highly concentrated in the gel than in the overlyingwater, with values of 8 and 13 mg/kg, respectively, compared to2.8 mg/L in CPY10 and 2.9 mg/L Pb in CPY14 for the bottom water,confirming its settling from the water column in form of anglesite.With the exception of Pb, none of the contaminant metals, Fe (2482and 3602 mg/kg), or Mn (137 and 179 mg/kg) were fixed in the gelbut were mobile in the gels pore water (water content 92 and91 wt.%). If metals were fixed in any form, the high hydraulic per-meability of the gel-like matter would have resulted in diffusionand downward migration of metals from the overlying lake waterinto the pore water and consequently in concentrations higherthan those in the lake water.

The rigid gel of CPY10(1020 and 2030 cm) had values formajor and trace elements similar to those of the overlying liquidgel (Fig. 4C) while rigid gel of CPY14 did not (Fig. 4B). ProfileCPY14 contained calcite below 18 cm and had a pH of 5.6. Directlyabove the calcite a layer with gypsum concretions formed,where up to 638 mg/kg Cu and 1908 mg/kg As had precipitated(Fig. 4B).

4.7. Trace metal and arsenic distribution in fine-grained sediments andsoils

The precipitation of Fe and Al was controlled by the pH increasewith depth. Iron was found at lower depth and lower pH than Al.This is in agreement with investigations at the micro scale for car-bonate grains reacting with AMD (Hammarstrom et al., 2003) andalso with field-size studies of oxic limestone drains (Rtting et al.,2008), which found a sequence of Fe- and Al-rich layers at the top,followed by gypsum with carbonate below. Controlling factors forFe and Al precipitation and trace element distribution were the pHgradient and Ca diffusion for gypsum formation (Hammarstromet al., 2003).

In CPY25, from the surface to 31 cm depth, concentrations in-creased from 197 to 5177 mg/kg Zn and from 78 to 808 mg/kgMn. This trend was observed in all profiles, where calcite was pres-ent. These values exceeded the concentration in bedrock (up to24 mg/kg Zn and up to 74 mg/kg Mn) by 215 times for Zn and 11times for Mn and implied that the bedrock was not an importantsource of Zn and Mn in the overlying layers. Modeling of mean lakewater in equilibrium with calcite resulted in SIs of 0.2 for rhodo-chrosite (MnCO3) and 0.2 for smithsonite (ZnCO3). Manganeseand Zn were probably fixed as carbonate minerals at depths wherecalcite was still oversaturated.

Adsorption capacity is controlled by the pH, and increases foroxyanions at low pH and for cations at high pH. Enrichment inAs was correlated to the presence of plant roots: it was enrichedup to 2402 mg/kg in CPY07(1518 cm) and up to 3328 mg/kg inCPY23(1015 cm) (Fig. 5C and D). The As concentrations were lesscorrelated to the TOC than to the visibly detected plant roots. In therhizosphere, Fe-hydroxides form plaques near the roots due todownwards O2(g) fluxes (Hansel et al., 2001), and As is adsorbedon the Fe plaques (Hansel et al., 2002; Voegelin et al., 2007). Insome cases, high As concentrations were directly correlated tothe presence of Fe due to adsorption on Fe-hydroxides, as inCPY06(1421 cm) with 9.0 wt.% Fe2O3 and 2157 mg/kg As (Dzom-bak and Morel, 1990). Lead, Cu and Cd were found where the pHincreased. Because Pb was always found above Cd in the profiles,a sorption sequence as found by Dzombak and Morel (1990) canbe assumed.

Trace element migration with depth can be inferred in the se-quence As < Pb Cu < Cd < Zn = Mn. This is in relative agreementwith the migration sequence of trace elements in carbonate soilsafter the Aznalcllar tailings spill, which showed low mobility forAs and Pb, moderate mobility for Cu, and high mobility for Cdand Zn (Dorronsoro et al., 2002). After the cessation of disposal,metals and particularly As (the element of most environmentalconcern in the studied case) may desorb and partly resorb duringpH changes and transformation processes within the gel-like andgypsum layers, e.g. to goethite (Acero et al., 2006).

4.8. Lessons learned

Whereas jarosite was found to precipitate in the sediments ofvarious pit lakes (e.g. Bozau et al., 2007; Triantafyllidis andSkarpelis, 2006), to the present authors knowledge Lake Yanamateis the first case where jarosite resulted in element cycling withinthe water column. Thus, the analysis of the lake surface waterwould not be sufficient to infer deposition of dissolved constitu-ents by jarosite precipitation.

The discharge of kerosene had no detectable influence on thesolubility of metals and As within the acidic lake water and didnot retain metals in the sediments. Possibly, it lowered the sus-pended solids within the water column, due to the formation ofOMA. However, this effect does not favor its discharge togetherwith acidic mine or process waters, because it has to be seen as a


remediation target on its own during closure planning. The organicrich gel-like matter precipitated from the water column had awater content of 90 wt.% and was not able to inhibit the infiltra-tion of AMD into the bedrock.

After 22 a of continuous interaction of highly acidic AMD withlimestone formations, the higher acidity at the lake bottom relativeto the lake surface water indicates that in Lake Yanamate AMDneutralization of lake water by limestone was limited.

However, water balance modeling for the years 19962002 sug-gested infiltration of 910 Mio. m3/a, exceeding discharge of 79Mio. m3/a (Volcan, 2002). Although a sequence (Fe- and Al-hydrox-ides followed by gypsum, covering calcite) similar to precipitatesin oxic limestone drains developed, these layers did not stop theinfiltration of lake water into the subsurface. Why the systemwas not suffering from complete clogging (as expected from theexperiment with hand specimens) needs further study, and couldbe related to the small grain size of the carbonate soil material,or possible retention of CO2 within the lake bottom sediments,increasing calcite dissolution (Arakaki and Mucci, 1995). Although,hydraulically conductive faults and different forms of karst exist inthe lake basin, contamination of surface waters in the area sur-rounding Lake Yanamate is not reported. Possibly, the hydraulicdepression of the open pit and its intersection by the Cerro de Pas-co-fault caused preferential groundwater flow from the lake to-wards the open pit.

5. Conclusions

At Lake Yanamate, a combination of hydrogeochemical, stableisotope, bulk chemical and mineralogical methods were used tostudy the controls on processes of the lake water chemistry andthe watersediment interaction at the lake bottom.

The mean lake water composition did not differ significantlyfrom the discharged AMD and was dominated by high Fe andSO4 concentrations and acidity. The lake showed an increase ofelectrical conductivity and the concentrations of most elementswith depth. Stable isotope results for d2H and d18O indicate a localorigin of discharged AMD. Trends of stable isotopes within the lakewater body showed the influence of evaporation for the upper10 m and the local infiltration of near-surface groundwater forthe eastern part of the lake. Below 15 m a stagnant zone withnear-equal values for stable isotopes and element concentrationshad developed. Modeling suggested that gypsum, anglesite, streng-ite and jarosite to some extent controlled the lake water chemistry.Anglesite caused a homogeneous concentration of 2.8 mg/L Pb inthe lake, whereas jarosite formed within the upper part of the lake,settled, and dissolved in the deeper part of the water column,resulting in downward element transport. This cycling was con-firmed by the presence of jarosite being limited to cores from67 m water depth. The low pH was the limiting factor in mineralprecipitation. Reaction of lake water with carbonate bedrock didnot cause lower acidity or higher pH in the bottom layer of thelake.

Discharged kerosene most likely caused the flocculation of Oilmineral agglomerates (OMA) that settled and formed a gel-likelayer on top of the lake bottom. Metal and As concentrations recal-culated to a wet base showed no enrichment within this materialcompared to the overlying water, thus the OMA did not adsorbthese elements. In the sediments below, a layer of gypsum withpH 11.5 was found on top of dissolving calcite with pH 5.37.In many samples the pore water pH was around 5.3, in accordanceto the modeled pH for lake water in equilibrium with calcite. Forthe sediments, a migration sequence of As < Pb Cu < Cd

Cerro de pasco

Documents

Transcript of Cerro de pasco