Hydrogeochemistry of Naturally Occurring Arsenic and Other Trace Elements in the Central Bolivian Altiplano

HYDROGEOCHEMISTRY OF NATURALLY OCCURRING ARSENIC AND OTHER TRACE ELEMENTS IN THE CENTRAL BOLIVIAN ALTIPLANO

Sources, mobility and drinking water quality

Mauricio Rodolfo Ormachea Muñoz

June 2015

TRITA-LWR PHD-2015:03 ISSN 1650-8602 ISBN 978-91-7595-580-3 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Cover illustration: The Marques River flanked by hills of volcanic rock in the southern part of Poopó Lake basin (Photograph: Mauricio Rodolfo Ormachea Muñoz ©, 2010).

© Mauricio Rodolfo Ormachea Muñoz 2015 PhD Thesis KTH-International Groundwater Arsenic Research Group Division of Land and Water Resources Engineering Department of Sustainable Development, Environmental Sciences and Engineering KTH Royal Institute of Technology SE-100 44 Stockholm, Sweden Reference to this publication should be writing as: Ormachea Muñoz, M. R. (2015) “Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano - Sources, mobility and drinking water quality”. PhD Thesis, TRITA LWR PHD 2015:03, 39 p.

ii Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

to my beloved son …

iii Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

iv Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

ACKNOWLEDGMENT This thesis is a part of the SIDA-Bolivia Project titled “Hydrogeochemistry of arsenic and heavy metals in the Poopó and Uru Uru Basin, Bolivian Altiplano”, I would like to acknowledge the financial support given by the Swedish International Development Cooperation Agency (Contribution number: 7500707606) for this project. I would also like to acknowledge the financial support given by the Agencia Española de Cooperación Internacional para el Desarrollo (AECID) (N° Ref. 11-CAP2_1282) for completion of this research. I would like to express my sincere gratitude to my supervisor Professor Prosun Bhattacharya at the Division of Land and Water Resources Engineering of the Sustainable Development, Environmental Science and Technology Department, for his support and guidance during entire time period of my studies and research at KTH. I express my most sincere thanks to Professor Jorge Quintanilla Aguirre, leader of the Hydrochemistry Group at the Instituto de Investigaciones Químicas IIQ-UMSA, Bolivia for giving me opportunity to participate in different research projects since 2001, initially as a chemical analyst and researcher, and later as PhD student. My gratitude to my co-supervisors Prof. Jochen Bundschuh, Dr. Ondra Sracek and Prof. Roger Thunvik, for their positive criticisms, discussions, and significant comments during the preparation of scientific papers based on my current research. Professor Jorge Quintanilla Aguirre and Dra. María E. García Moreno, both SIDA project coordinators at UMSA in Bolivia, supported me in many different ways during the progress of my doctoral studies, I am grateful to both of them for their assistance and guidance. I owe a lot of credit to all people at DIPGIS-UMSA for administrative assistance, especially, thanks to MSc. Ignacio Chirico, SIDA-UMSA coordinator. I would like to acknowledge the help provided by Ann Fylkner and Bertil Nilsson for laboratory analyses at the Division of Land and Water Resources Engineering KTH, I am also grateful to Carl-Magnus Mörth, Department of Geological Sciences, Stockholm University for performing analyses of water samples collected from Bolivia through ICP- OES. My gratitude also goes to Md. Annaduzzaman for helping me formatting the thesis. My gratitude to people of the Instituto Geológico y Minero de España for all the good discussions and feedback during the field trip and skype meetings. I am very grateful to all colleagues and friends at KTH and UMSA in Sweden and in Bolivia for their moral support during my research at KTH. My special thanks to all of my family members, who supported and encouraged me to complete this research. Thanks for their understanding, without their unconditional support completion of my Doctoral Thesis was not possible. I also would like to thank people of the Bolivian Altiplano who allowed me to enter into their communities and homes to collect water samples. I have a strong commitment to these people to help them improve the quality of their drinking water, and thereby improve the quality of their lives.

Mauricio Ormachea Muñoz Stockholm, June 2015

v Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

vi Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

LIST OF APPENDED PAPERS AND MY CONTRIBUTIONS This thesis is based on the following four papers, which are referred as corresponding Roman numerals throughout the text and are attached in the Appendix. Paper I Ramos Ramos, O.E., Cáceres, L.F., Ormachea Muñoz, M., Bhattacharya, P., Quino, I., Quintanilla, J., Ondra, S., Thunvik, R., Bundschuh, J., García, M.E. 2012. Sources and behavior of arsenic and trace elements in groundwater and surface water in the Poopó Lake Basin, Bolivian Altiplano. Environmental Earth Sciences 66:793–807. DOI: 10.1007/s12665-011-1288-1 (Reprinted with permission from Springer). I participated in the sampling campaign, field work, geochemical modeling, data analysis, interpretation and writing. Paper II Ormachea Muñoz, M., Bhattacharya, P., Sracek, O., Ramos Ramos, O., Quintanilla Aguirre, J., Bundschuh, J., Maity J. 2015. Arsenic and other trace elements in thermal springs and in cold waters from drinking water wells on the Bolivian Altiplano. Journal of South American Earth Sciences 60: 10–20. DOI:10.1016/j.jsames.2015.02.006 (Reprinted with permission from Elsevier). I participated in the scientific planning, field trip, samples collection and analyses, data analyses, interpretations, geochemical modeling and writing. Paper III Ormachea Muñoz, M., Wern, H., Johnsson, F., Bhattacharya, P., Sracek, O., Thunvik, R., Quintanilla, J., Bundschuh, J. 2013. Geogenic arsenic and other trace elements in the shallow hydrogeologic system of Southern Poopó Basin, Bolivian Altiplano. Journal of Hazardous Materials 262 924– 940. DOI: 10.1016/j.jhazmat.2013.06.078 (Reprinted with permission from Elsevier). I participated in project design, field trip, water and sediments samples collection and analyses, data analyses, interpretations, geochemical modeling and writing. Paper IV Ormachea Muñoz, M., García Aróstegui, J.L., García Moreno, M.E., Kohfahl, C., Quintanilla Aguirre, J., Hornero, J., Bhattacharya, P. Geochemistry of naturally occurring arsenic in groundwater and surface-water in the southern part of the Poopó Lake basin, Bolivian Altiplano. Manuscript submitted to Journal of Groundwater for Sustainable Development (In review). I participated in project design, field trips, samples collection and analyses, data analyses, interpretations and writing. LIST OF PAPERS NOT APPENDED IN THE THESIS Paper V Quintanilla, J., Ramos Ramos, O.E., Ormachea Muñoz, M.R., García, M.E., Medina, H., Thunvik, R. & Bhattacharya, P. 2009 Arsenic contamination, speciation and environmental consequences in the Bolivian plateau. In: J. Bundschuh, M.A. Armienta, P. Birkle, P. Bhattacharya, J. Matschullat & A.B. Mukherjee (eds.): Natural Arsenic in Groundwater of Latin America ― Occurrence, health impact and remediation. Interdisciplinary Book Series: Arsenic in the Environment Volume 1, J. Bundschuh & P. Bhattacharya (Series Editors), CRC Press/Balkema, Leiden, The Netherlands, pp. 91-100. ISBN: 978-0-415-40771-7.

vii Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Paper VI López, D., Bundschuh, J., Birkle, P., Armienta, M.A., Cumbal, L., Sracek, O., Cornejo, L., Ormachea Muñoz, M.R. 2012. Arsenic in volcanic geothermal fluids of Latin America, Science of the Total Environment, 429 55-75.

viii Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

TABLE OF CONTENTS Acknowledgment ...... v List of Appended Papers and my Contributions ...... vii List of Papers not Appended in the Thesis ...... vii Table of Contents ...... ix Abstract ...... 1 1. Introduction ...... 2 1.1. Background ...... 2 1.2. Arsenic in the environment ...... 4 1.3. Geochemistry and mobility of As ...... 5 1.4. Health problems related to As exposure ...... 6 1.5. Drinking water quality compliance with Bolivian regulation and WHO guidelines ...... 7 1.6. Uses of water resources in rural Bolivian Altiplano ...... 7 1.7. Rationale ...... 8 2. Research Objectives...... 10 2.1. General objective ...... 10 2.2. Specific objectives ...... 10 3. Study Area Description ...... 10 3.1. Physiography and climate ...... 10 3.2. Geology and hydrogeology...... 10 4. Materials and Methods ...... 12 4.1. Field investigations ...... 12 4.1.1. Surface and groundwater sampling ...... 12 4.1.2. Sediment sampling ...... 14 4.1.3. Core-sediment sampling ...... 14 4.1.4. Rock sampling ...... 14 4.2. Laboratory analysis ...... 14 4.2.1. Water analysis ...... 14 4.2.2. Geochemical characterization of sediments from geothermal springs ...... 14 4.2.3. Geochemical characterization of core sediments...... 14 4.2.4. Rock analyses ...... 15 4.3. Geochemical modelling and data analyses ...... 15 4.4. Quality control ...... 15 4.4.1. Water samples ...... 15 4.4.2. Rock and sediment samples ...... 15 5. Results ...... 15 5.1. Hydrochemistry of groundwater and surface water within the Poopó Lake basin ...... 15 5.1.1. Groundwater ...... 15 5.1.2. Surface water ...... 16 5.1.3. Geothermal springs ...... 17 5.2. Hydrochemistry of groundwater and surface water in the southern part of Poopó Lake basin ...... 18 5.2.1. Groundwater ...... 18 5.2.2. Surface water ...... 19 5.3. Geochemical modelling ...... 20

ix Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

5.4. Geochemical investigations ...... 21 5.4.1. Arsenic and TEs in sediments from geothermal springs ...... 21 5.4.2. Arsenic and TEs in core-sediments ...... 21 5.4.3. Arsenic and TEs in rocks ...... 22 6. Discussion ...... 22 6.1. Origin of surface and groundwater composition ...... 22 6.2. Spatial distribution of As and other TEs ...... 24 6.2.1. Groundwater within the Poopó Lake basin ...... 24 6.2.2. Groundwater and surface water in the southern part of Poopó Lake basin ...... 25 6.3. Seasonal variations ...... 27 6.4. Sources of As and main mechanisms for As mobilization ...... 29 6.4.1. Geothermal springs ...... 29 6.4.2. Rock types ...... 29 6.4.3. Sediments...... 31 6.5. Drinking water quality ...... 32 7. Conclusions ...... 32 8. Future Work ...... 33 9. References ...... 34

x Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

ABSTRACT The Bolivian Altiplano (BA) is a high plateau located in the western part of the country at an altitude of 3,600 to 3,900 meters above sea level and is bordered by the Eastern and Western Cordillera. Within the BA there is a large endorheic hydrologic system linking the Titicaca Lake in the north the Desaguadero River, lakes Uru-Uru and Poopó in the central part; and the Lacajahuira River and Coipasa and Uyuni salt pans in the south. Several mineralized areas, especially in the Eastern Cordillera, have been intensively exploited for centuries for the extraction of silver, gold, and tin from polymetallic sulfide ore deposits. Presently many urban centers are also contributing for an extensive contamination in localized areas; especially the Poopó Lake and some rivers are affected by high loads of wastewater and solid waste, in addition to the release of heavy metals and arsenic (As) due to acid mine drainage. The presence of As in the BA was known to be related to mining only, but recent studies revealed that As of geogenic origin also contributing to the elevated concentration of As in surface and groundwater. The Poopó Lake basin is characterized by a semiarid climate. Geologic formations predominantly are of volcanic origin and groundwater flow is sluggish in nature. These environmental settings have generated substantially elevated concentrations of geogenic As and other trace elements in surface and groundwater. Both surface and groundwater used for drinking water have high concentrations of As that by far exceed the World Health Organization (WHO) guideline. The overall objective of the present study has been focused on the determination of the sources and principal mechanisms for mobilization of geogenic As into surface and groundwater of the Poopó Lake basin area. More specifically, this study has determined the spatial distribution and the extent of As contamination in surface and groundwater; chemical composition of surface and groundwater, rock and sediment; major geochemical mechanisms for As mobilization from solid phase to aqueous phases. This study also made an assessment of drinking water quality in rural areas within the Poopó Lake basin. Arsenic concentration exceeded the WHO guideline and national regulations for drinking water of 10 µg/L in 85% of the samples collected from the area around the Poopó Lake (n=27) and 90% of the samples from the southern part of the lake basin (n=42). Groundwater samples collected from drinking water wells had As concentrations up to 623 µg/L, while samples collected from piezometers had even higher up to 3,497 µg/L. Highest concentration in river water samples was observed 117 µg/L. Alkaline nature of water (median pH 8.3 for groundwater and 9.0 for surface water), predominance of Na-Cl-

HCO3 water type and elevated Eh reflecting oxidized character has been revealed by As(V) as the major species in As speciation. Different rock types were analyzed for their As content and the highest concentration of 27 mg/kg was found in a coral limestone sample. In evaporate it was 13 mg and 11 mg As/kg was measured in calcareous sandstone. Elevated concentration of As was also observed in sediment cores collected from two drilling sites; 51 mg/kg in Condo K and 36 mg/kg in Quillacas. Physical and chemical weathering of volcanic rocks, limestone, carbonates and plagioclase minerals enhance the + - supply of Na and HCO3 into solution and as a consequence pH and alkalinity increase, which in turn, favor As desorption from solid mineral surfaces (especially Fe(III) oxyhydroxides) and therefore dissolved As in water is increased.

Keywords: Arsenic; Anthropogenic; Aquifer; Bolivian Altiplano; Boron; Drinking water quality; Geochemistry; Geogenic; Groundwater; Lithium; Trace Elements.

1 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

1. INTRODUCTION into the Coipasa salt pan (Montes de Oca, 1997). Several activities are being carried out 1.1. Background in the BA; the most important are mining, The presence of arsenic (As) in drinking agriculture, and cattle rising. As per the latest water is a major concern to human health census report within the BA are living due to its neurotoxicity and potential to cau- 3,451,349 inhabitants (INE, 2005). se cardiovascular diseases and different types The Eastern and Western Cordillera are rich of cancer, of which skin and bladder cancers in mineral resources; several mining areas, are the most common (Kapaj et al., 2006; especially in the Eastern Cordillera, have Chen, 2010). Globally, an estimated 200 been exploited intensively for centuries from million people are directly affected by As pre-colonial period to the present time for contamination; a number which is constantly silver, gold, and tin from polymetallic (Fe, increasing as new contaminated areas are Cu, Zn and Pb) sulfide ore deposits, causing being discovered (Bundschuh and Bhattach- severe environmental damages to the region arya, 2012). Naturally occurring As (geogenic (García et al., 2005). As) in water, mainly in groundwater, is attri- Sustained complaints of villagers about the buted to natural geochemical processes but degradation of the environment and non- also due to many labor activities such as compliance with national regulations led to mining (Smedley and Kinniburgh, 2002). initiation of several research projects in the Concentrations of As in drinking water abo- BA. Environmental monitoring programs ve the World Health Organization guideline have been undertaken for the assessment of (WHO) (10 µg/L) (WHO, 2011) have been environmental conditions in the region. documented across the world, including Most of the studies point out mining and countries such as India, Bangladesh, China, urban centers as the main anthropogenic Vietnam, Taiwan, Spain, Portugal and sources responsible for environmental pollu- Hungary (Bhattacharya et al., 1997; Smedley tion with large discharges of urban wastes in and Kinniburgh, 2002; Bhattacharya et al., addition to tailings and waste rocks. Also, 2002a, 2002b; Bhattacharya et al., 2007; agro-industrial products, especially vegeta- Bhattacharya et al., 2011), as well as nume- bles like potatoes, onion and carrots are rous countries in Latin America such as Me- affected by pollution of soils with heavy xico, Nicaragua, Ecuador, Chile, Argentina, metals and As through the deposition of Peru, Brazil and Uruguay (Bhattacharya et atmospheric particulate matter coming from al., 2006; Bundschuh, 2008; Cumbal, 2009; smelters in suburbs of the El Alto and Bundschuh, 2009; Bundschuh, et al., 2012). Oruro cities (Fig. 1) (Quintanilla et al., 2009; In Bolivia, however, only few studies have Tapia et al., 2012; Rötting et al., 2013). been conducted regarding the presence of The Poopó Lake basin (Fig. 1) located As in the environment; most noteworthy within the BA is the area which has received investigations were carried out mainly in the most attention for environmental assess- Bolivian Altiplano (BA). The BA is a high ments due to extensive mining activities plateau geographically located in the western carried out especially in the northern and part of the country between 3,600 to 3,900 northeastern parts of the basin where acid meters above sea level and bordered by the mine drainage (AMD), acid rock drainage Eastern and Western Cordillera. Within the (ARD), weathering of waste rocks and BA lies a large endorheic hydrologic system tailings together with wastewater and urban (TDPS system) comprising the Titicaca Lake waste have left an extensive contamination in the north, which drains through the of river banks, surface water and soils with Desaguadero River and stretches 380 km high loads of organic matter, dissolved over the plain before discharging into the minerals, heavy metals and of course, As lakes Uru-Uru and Poopó (Fig. 1). Water (PPO 1992; PPO 1996a, 1996b, 1996c; from the Poopó Lake drains to the south García et al., 2006). into Lacajahuira River, which finally drains

2 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Titicaca Lake

El Alto

Desaguadero River

Poopó Lake

Salt flats

Elevation m above sea level

Figure 1. The Bolivian Altiplano, TDPS System and Poopó Lake basin.

Quintanilla et al. (1995) reported concentra- 245 µg/L while water of a stream affected tions As of up to 800 µg/L in the Desagua- by both AMD and geothermal activity dero River located in the Central BA (Fig. 1), contained a maximum concentration of while in the same river, the Oruro Pilot 11,140 µg/L. The occurrence of As in the Project (PPO, 1996b) reported As concen- BA hydrological systems is thus attributable trations of 588-1,180 µg/L (average: 740 to both anthropogenic and geogenic sources. µg/L). Banks et al. (2004) assessed the surfa- Geogenic As in drinking water was first been ce water drainage in the catchment areas of documented during the evaluation of the the salt pans of Coipasa and Uyuni, located environmental status in the mining district to the southwestern part of the BA (Fig. 1) of Oruro, where the project “Pilot Project where concentration of As was found to be Oruro” (PPO, 1996a) reported high concen- as high as 4,600 µg/L (median 34 µg/L; trations of As in drinking water wells from n=297). The bulk of As is geogenic, released shallow aquifers in the southern and western through leaching of volcanic rocks and their part of the Poopó Lake basin (Fig. 1), ran- weathering products, and local geothermal ging from 33 to 299 µg/L; average 181 µg/L springs within the Western and Eastern Cor- (n=5) in Quillacas and Sevaruyo, while in dillera (PPO 1996a; SERGEOMIN, 1999). Toledo, concentrations were in the range of Quintanilla et al. (2009) reported elevated 121-957 µg/L; average 463 µg/L (n=5) concentrations of As from geogenic and exceeding by far the international guideline anthropogenic sources in both groundwater value for drinking water (10 µg/L) (WHO, as well as surface water within the Poopó 2011). At a first glance, results suggested Lake basin (Fig. 1). Groundwater in this area that the presence of high concentrations of contained As concentrations of up to As in shallow aquifers might be due to an

3 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03 enriched geological background in the As chemistry and its behavior in different region; thus the PPO report strongly natural geological settings; thus in the next recommended more extensive and conclusi- sections a theoretical background will be ve investigations to determine natural sour- provided to better understand the geo- ces of As and its health impacts on the chemistry of As. population. 1.2. Arsenic in the environment After eight years of the publication of the Arsenic (As, atomic number 33, atomic first report on geogenic As in drinking water weight 74.922) is a naturally occurring wells by PPO (PPO, 1996a), new investi- element which is widely disseminated in the gations were initiated by the Chemical atmosphere, hydrosphere and biosphere, it is Research Institute of the Universidad Mayor the 20th most common element in the earth’s de San Andrés (UMSA) in La Paz, Bolivia, crust; approx. 5 mg/kg (Tanaka, 1988). through the Sida-Bolivia Cooperation Pro- While a large quantity of As in the environ- gram starting from 2000. In this investi- ment comes from natural sources (weathe- gation, the assessment of drinking water ring, biological activity, geothermal activity, quality was carried out focusing on the volcanic emissions); there also exist impor- occurrence of As and other trace elements tant contribution from anthropogenic activi- (TEs) in rivers and shallow aquifers in ties such as industrial processes like mining, Poopó Lake basin (Fig. 1). Hermansson and smelting of metals, use of herbicides and Karlsson (2004) reported As concentrations wood preservatives (Smedley and Kinni- ranging from below detection levels burgh, 2002; Welch and Stollenwerk, 2003; (1.5 µg/L) up to 245 µg/L (average 47 µg/L; Henke, 2009). n=23) in drinking water wells located in communities around the Poopó Lake. This The oxidation states of As are: As(-III), study mentioned the heterogeneous spatial As(0), As(III) and As(V). In natural waters, distribution of As contamination in the area inorganic As occurs as dissolved species for- and also confirmed that the most impacted ming oxyanions (negatively charged ions area with high concentrations of geogenic combined with oxygen) typically in the form As in drinking water is located in the of trivalent arsenite [As(III)] or pentavalent southern part of Poopó Lake basin (Fig. 1). arsenate [As(V)]. Arsenite occurs as H3AsO3 Quino (2006) found low As concentrations and its corresponding protolithic derivatives + - 2- 3- of up to 14 µg/L in groundwater from (H4AsO3 , H2AsO3 , HAsO3 , AsO3 , pka1 = shallow aquifers in the northeastern part of 2.3; pka2=9.2 and pka3=12.7), while arsenate Poopó Lake basin while Van Den Bergh et occurs as H3AsO4 and its corresponding - 2- al. (2010) reported concentrations of As up dissociation products (H2AsO4 , HAsO4 , 3- to 964 µg/L (average of 63.6 µg/L) in drin- AsO4 , pka1=2.3; pka2=6.8; pka3=11.6). king water wells from communities located Organic forms of As occur, mostly due to north and northeast of Poopó Lake (Fig. 1). biological activity, in surface waters or where In a recent investigation, Tapia et al. (2012) waters are significantly impacted by indus- studied lake sediments within the Poopó trial pollution, but are rarely quantitatively Lake basin, which revealed that the natural important (Smedley and Kinniburgh, 2002; geochemical background of As in the basin Henke, 2009). is significantly enriched in comparison to The speciation of As in aquatic systems is upper continental crust; they also suggested controlled by redox potential (Eh) and pH. that mixed As sources could explain the The distributions of different As species in heterogeneity on the distribution of As. the aqueous system is presented in the Eh- Investigations of geogenic As on the envi- pH diagram (Fig. 2). Arsenate is the domi- ronment within the BA are in its initial nant specie under oxic conditions and is stages, addressing the problem of As in found in the common pH range of most drinking water requires different approaches natural surface and groundwaters as the oxy- - 2- which take into account the knowledge of anions H2AsO4 at pH < 6.8) and HAsO4 at

4 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

for the occurrence of high concentrations of As in groundwater and surface water (Smedley and Kinniburgh, 2002). Geogenic sources of As are numerous; As is found in various geologic materials such as rocks, minerals, soils and sediments and also originates from geothermal springs. Geo- thermal waters rich in As are associated with areas of active volcanism; the mixing of As- rich geothermal groundwater with cold aquifers was identified as one of the main environmental problem in As contamination, As-rich surface waters are also found in rivers and lakes near the hot spring discharges (Birkle and Bundschuh, 2007; Figure 2. Eh – pH diagram of aqueous Cumbal et al., 2009; López et al., 2009; As species in water at 25°C and 1 bar Bundschuh and Maity, 2015). total pressure (modified from Panagio- Average concentration of As in the earth´s taras et al., 2012). crust is 5 mg/kg (Tanaka, 1988; Smedley and Kinniburgh, 2002; Welch and Stollenwerk, pH > 6.8). Under reducing conditions, 2003). In igneous and metamorphic rocks arsenite is the predominant species, being concentration of As is low (1-5 mg/kg), but the principal form in most reduced natural it is slightly higher in soils and sedimentary rocks (5-10 mg/kg) (Smedley and Kinni- waters the neutral specie H3AsO3 at pH < 9.2 (Mukherjee and Bhattacharya, 2001; burgh, 2002). Arsenic bearing minerals are Smedley and Kinniburgh, 2002; Appelo and found in crustal rocks out of which most Postma, 2005). common are arsenopyrite (FeAsS) and As bearing pyrite (FeS ); As also appears within Generally, in surface water arsenate is predo- 2 some silicate and carbonate rocks and as minant over arsenite. Average concentration sorbed species on secondary As minerals of As in river waters is mostly below such as oxides and oxy-hydroxides of Fe, 0.8 µg/L; however, it can be higher depen- Mn and Al. Major geologic sources of As ding on many factors such as groundwater include alluvial sediments, mineralized sedi- withdrawal, groundwater recharge, mixing mentary and metasedimentary rocks as well with water from geothermal springs, weather as rocks of volcanic origin and related conditions, leakage from mineralized zones, deposits of their weathering products mining industry, urban waste and industrial (Bhattacharya et al., 2002a; Smedley and waste. In lakes, average concentration of As Kinniburgh, 2002). in water is similar to those in rivers because factors are almost the same. 1.3. Geochemistry and mobility of As In groundwater, both As(III) as As(V) can Behavior of As in natural environmental sys- be found since As species are dependent on tems is complex; multiple geochemical pro- the amount of dissolved oxygen, pH, redox cesses drive mobilization of As into ground- conditions and biological activity. Concen- water and surface water. trations of As in groundwater vary in a wide The concentration of As in aquatic environ- range from <0.5 to 5,000 µg/L. Although in ments depends on the available As sources, localized areas different anthropogenic acti- liquid properties (pH, Eh, ionic strength, veties are responsible for water pollution etc.), solid/liquid interaction processes and resulting in high concentrations of As in the presence of other ions (ions competing water, in the majority of regions natural with As for adsorption sites). The principal geochemical processes are principal factors processes for As mobilization include disso-

5 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03 lution/precipitation, adsorption/desorption, 1.4. Health problems related to As ion exchange, redox and microbiological exposure transformations. Weathering of minerals in Human beings may be exposed to As the crustal rocks, especially oxidation of through inhalation, dermal absorption, and primary sulfide minerals is the main source ingestion of contaminated food, water, and of As in soil and sediments from where soil. As(III) and As(V) in ingested water are naturally occurring As is generally leached well absorbed from the gastrointestinal tract. into the groundwater system by the interact- Arsenite has been considered the more toxic tion of water with rocks, minerals soil and species as compared to arsenate; however, sediments in different geological, hydro- some studies have shown that most ingested geological and geochemical environments. arsenate is reduced to arsenite. The biotrans- Depending on the geochemical environ- formation of inorganic As [As(V) and ment, iron oxides and oxyhydroxides can act As(III)] within the human body, takes place as both sources and sinks for dissolved As. through a multistep reaction mediated by a Iron oxides and oxyhydroxides are conside- series of enzymes; therefore, the inges- red as principal adsorbents of As(V) in many tion and exposure of both species may result sand and silt aquifers (Dzombak and Morel in similar toxicological effects (Welch and 1990; Bhattacharya et al., 2002b; Smedley Stollenwerk, 2003; Aposhian et al., 2004; and Kinniburgh, 2002; Mitsunobu et al., Henke K., 2009). Arsenic-induced skin 2008). lesions may include keratosis (including Mobility of As in natural environment is hyperkeratosis and parakeratosis), arsenical predominantly governed by the redox condi- pigmentation, melanosis, and squamous cell tions. Sracek et al. (2004) classified the beha- carcinoma. The lesions are normally most vior of As in aquifers from three distinct pronounced on the feet and hands, although depth specific redox-zones: they can occur on the trunk and other areas i) shallow, oxidized zone with dissolved of the extremities. Epidemiological studies oxygen, in which Fe(III)-oxyhydroxides are linked As exposure to skin cancer and stable and constrains the mobility of As(V) various other internal cancers, including through adsorption at high pH; angiosarcoma of the liver, lung cancer and ii) intermediate depth zone with moderately bladder cancer. reduced environment, devoid of dissolved Besides the carcinogenic effects, chronic oxygen, accompanied by release of As due to exposure of As also affects gastrointestinal reductive dissolution of Fe(III)-oxyhydr- system in humans. Other effects of As expo- oxides; and sure include noncirrhotic portal hyperten- 2- sion, hepatic or splenic enlargement, internal iii) deep reducing zone, where SO4 is reduced followed by co-precipitation of As cancers such as hepatocellular carcinoma in form of secondary sulfides. However, As and liver angiosarcoma. Recent studies have is mobile if the concentration of dissolved suggested an association between As expo- SO 2- is low. sure and an increased risk for a variety of 4 non-cancer effects, which are manifested as Ion competition may also plays an important peripheral vascular disease, cardiovascular role in mobilizing As adsorbed onto reactive disease, diabetes, neurological effects, chro- surfaces. For mobilizing adsorbed As into nic lung diseases, diminished hearing and groundwater, the most common ions in 3– 2– cerebrovascular disease. Thus, the health order of importance are PO4 , SO4 , – effects of As are related to multiple organs HCO3 and H4SiO4; the dissolved ions in human body linked with extensive system competing directly for specific adsorption pathology (Centeno et al., 2000; Kapaj et al., sites or indirectly through their influence on 2006; Henke, 2009). the electrostatic charge of the reactive surfaces (Welch and Stollenwerk, 2003).

6 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

1.5. Drinking water quality compli- Table 1. Comparison of selected ance with Bolivian regulation and parameters between World Health WHO guidelines Organization (WHO, 2008) and As result of severe chronic toxicological Bolivian standard for drinking water effects caused by human exposure to As (NB 512, 2005). through drinking water, health organizations WHO NB 512 and regulatory authorities perceived the Parameter Unit Guideline Maximum value acceptable severity of the problem and started to intro- Al mg/L 0.2a 0.1 duce or lower the permissible limits for As in drinking water. In 1993, the WHO decre- As μg/L 10 10 ased the provisional guideline for As in B μg/L 500 300 drinking water from 50 to 10 µg/L; since Ca mg/L - 200 Cd μg/L 3 5 then, many countries around the world have − a introduced in their national regulations the Cl mg/L 250 250 WHO guideline value for As content in Cr μg/L 50 50 Cu μg/L 2,000 1,000 drinking water (WHO, 1993; WHO, 2003). − In Bolivia, especially in urban centers, water F mg/L 1.5 1.5 a supply is performed mainly by public utilities Fe μg/L 300 300 usually via a system of dams, pumps and Mg mg/L - 150 pipes. Public or private entities providing Mn μg/L 400 100 drinking water services are obligated by law Na mg/L - 200 to guarantee the quality of water through the Ni μg/L 20 50 − implementation of the Bolivian National NO3 mg/L 50 45 Regulation for Control the Quality of Water Pb μg/L 10 10 for Human Consumption (NB 512, 2005); Se μg/L 10 10 2− local governments and municipalities should SO4 mg/L - 400 comply with the Bolivian national regulation. U μg/L 15 - Control parameters for drinking water qua- Zn μg/L 1,000 a 5,000 lity in Bolivian regulation depend on TDS mg/L 1,000a 1,000 technical and economic feasibility and are pH - 6.5-8.0a 6.5-9.0 gathered in four groups: minimal control aRecommendation based on aesthetic parameters, basic control parameters, com- considerations (taste and color). plementary control parameters and especial control parameters; As is grouped within the are private dug wells, communal wells, complementary control parameters and at a springs and streams; home owners and local maximum permissible limit of 10 µg/L authorities are responsible for the drinking (NB 512, 2005). However, in Bolivia, there water quality of private and communal wells are very few laboratories with adequate but due to the extreme poverty in this instrumentation and trained personnel to region, drinking water wells are never tested quantitatively determine concentrations of for water quality parameters, which is an As at that low concentration level, which alarming concern for public health. Water is makes the assessment of drinking water qua- a potential hazard in this area because it may lity very expensive and complicated. Table 1 contain high concentrations of As and other shows a comparison between Bolivian regu- toxic contaminants. lation (NB 512, 2005) and international guidelines (WHO, 2008) for As and other selec- 1.6. Uses of water resources in rural ted parameters. Bolivian Altiplano In rural areas of the BA, public utilities and Bolivia currently has a relatively stable infrastructure for drinking water supply are economy due to its natural resources (natural seldom existing; e.g., within the Poopó Lake gas and mineral resources) that have contri- basin, common sources for drinking water buted to a decline in poverty rate in recent

7 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03 years. However, Bolivia is one of South within their households because surface America’s poorest countries; gaps exist water frequently dries up during dry season. between different geographical regions and 1.7. Rationale between urban and rural areas. The highest level of poverty is widespread and deeply As discussed in the background section, entrenched in rural areas of the BA, which urban centers and mining industry are affects quality of life of the majority of the principal anthropogenic activities response- population. Most of inhabitants of this area, ble for the environmental degradation of the especially those in semi-arid areas such as BA (García et al., 2005; Quintanilla et al., the Poopó Lake basin, have no access to 2009; Tapia et al., 2012; Rötting et al., 2013); basic sanitation services or to safe drinking e.g., the Poopó Lake basin has suffered water supply (TWB; UNICEF). severe environmental impacts especially to the north and north-east areas where solid Groundwater in large parts of the BA, waste and wastewater discharge together especially within the Poopó Lake basin, is with tailings and waste rocks have polluted the principal water source and is being water bodies, soils and air in localized areas extracted for household and agricultural around mines and major cities (PPO 1992, purposes from shallow wells. Water extrac- 1996a, 1996b, 1996c; García et al., 2006; ted from private and communal wells is used Tapia et al., 2012; Rötting et al., 2013). mainly for drinking, cooking and washing. Private wells are located within households For long time As occurrence in the ecosys- while communal wells are usually located in tem has been always related to the mining, the central part of villages, and depending but recent investigations carried out during on the size of the population, more than one last ten years in areas not impacted by communal well may be present in one mining revealed that an enriched geological village. background is responsible for elevated concentrations of As in different environ- Usually communal wells are deeper than mental matrices such as soils, rocks, sedi- private wells and thus dry up less frequently. ments, groundwater and surface water which Wells are excavated by hand shovel to the is a big concern as they are principal sources depth below the water table; the well in a for drinking water (Quintanilla et al., 1995; few cases is lined with stones or other PPO, 1996a; SERGOEMIN, 1999; Banks et material to prevent collapse, and top is cove- al., 2004; Hermansson and Karlsson, 2004; red with a cap of wood, stone, or concrete Van Den Bergh et al., 2010; Tapia et al., (Fig. 3). These types of wells are shallow and 2012). lack of continuous casing makes them sub- ject to contamination from nearby surface Although geogenic As has been studied in sources. Many such wells dry up towards the many countries all around the world, there is end of dry season. Common depths of both no comprehensive study on the contamina- private and communal wells are between tion of surface and groundwater from an- two to nine meters and diameters between thropogenic or geogenic sources in Bolivia, one to two meters; the most common way as well as their impact on the population. that people extract water from a well is The knowledge about the extent of As pulling a rope tied to a bucket (Fig. 3). distribution in surface and groundwater, the Surface water from rivers and streams in main geological sources and the principal some areas of the BA is used for irrigation geochemical mechanisms for As mobiliza- and drinking water for livestock, although in tion still has not been fully understood. some locations they are also used as water More extensive investigations are required in source for human consumption. Sometimes order to determine the extent of geogenic As surface water in the headwaters is diverted contamination, the main geological sources, from the river in galleries upstream of villa- and the principal geochemical processes for ges for using it as drinking water; however, As mobilization. villagers near rivers also have private wells

8 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Thus, the primary scientific questions ad- and ground water affected by rock-, dressed within this thesis are: sediment-water interactions, leading to  What is the extent of As contamination their pollution with high concentrations in the Poopó Lake basin within the BA? of natural As and other toxic TEs?  What are the sources of As in natural  How all these processes affect the quality waters? of drinking water?  What are the different geochemical Hence, quantitative knowledge of governing processes which govern the mobilization geochemical processes is essential not only of As from geological sources into to understand the principal mechanisms for groundwater and surface water? As release, but also to plan possible miti-  How is the chemical composition and gation actions for the remediation of natural physical characteristics of surface water excess contaminant concentrations.

Figure 3. Drinking water wells: a) Private well with concrete protection; b) wall lined with stones (seen from above); c) wall unlined (seen from above); d) communal well with concrete protection.

9 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

2. RESEARCH OBJECTIVES 3. STUDY AREA DESCRIPTION 2.1. General objective 3.1. Physiography and climate The overall objective of this study was to The Poopó Lake basin is located in the improve the understanding and to enhance central BA being part of the TDPS system the current knowledge on the occurrence of (Fig. 1). The lake is located at 3,600 meters As and other TEs in groundwater and above sea level and has a maximum length surface water with special emphasis on geo- of 90 km, a maximum width of 53 km and genic As in BA. In this context, the covers an area that varies – depending on identification of principal geogenic sources the season – from 2,650 to 4,200 km2 was highly significant to determine the (Quintanilla, 1994). The Poopó Lake basin is different geochemical mechanisms that rele- characterized by a cold semi-arid climate ase and mobilize As into groundwater and with an annual average temperature of 7ºC surface water. Furthermore, this study had as with a significant daily variation. Most preci- another major objective to perform the pitation falls as rain during the wet season physicochemical characterization and assess- from November to March, the precipitation ment of drinking water quality focused on is higher closer to the Eastern Cordillera geogenic As in the rural area of the Poopó (Huaranca Olivera and Neumann Redling, Lake basin located within the BA. 2000). The mean annual rainfall for a period of 42 years (1960-2002) is 372 mm (Pillco 2.2. Specific objectives and Bengtsson, 2006). Extreme weather situ- The specific objectives of the study were to: ations are frequent with both droughts and i) Investigate the extent and distribution of floods, depending on the natural variation of As and other TEs and to determine the precipitation further upstream in the basin hydrogeochemical characteristics of sur- (García, 2006). Erosion of the top soil due face and groundwater in the Poopó Lake to overgrazing and deforestation is an urgent basin (Paper I & II), and in a more problem in the BA (Van Damme, 2002). As specific area located to the southern part discussed in the background section, in the of Poopó Lake basin (Paper III & IV). northern and northeastern Poopó Lake ii) Determine the hydrogeochemical charac- basin, urban waste and intense mining have teristics of geothermal springs, and their been polluting localized areas around large potential role as sources for high As le- villages, industries and mines, while in the vels in groundwater within the Poopó western and southern parts of the basin, Lake basin (Paper II). cattle rising and agricultural activities have iii) Identify the major geogenic sources and managed to maintain healthy environmental to determine the principal geochemical conditions (Fig. 4). processes governing the release and mobilization of As into surface and ground- 3.2. Geology and hydrogeology water in the Poopó Lake basin (Paper I & The Poopó Lake basin is located between II), and more specifically in an area loca- the Eastern and the Western Cordillera of ted to the southern part of Poopó Lake the Andes, which have a different geological basin (Paper III & IV). evolutionary history (YPFB-GEOBOL, iv) Investigate the effect of seasonal varia- 1996). The Eastern Cordillera is composed tion on the spatial distribution of geo- of Paleozoic and Mesozoic rocks, which genic As and other TEs in groundwater were deposited on the Precambrian base- and surface water and furthermore to ment and deformed by multiple orogenic assess the drinking water quality with cycles (Paleozoic to Cenozoic); the principal special emphasis on geogenic As in rocks are Ordovician limestones, quartzites, drinking-water wells in the southern part slates and schists with low angle faults. The of Poopó Lake basin (Paper III & IV). Western Cordillera is composed of volcanic material accumulated in late Tertiary and early Quaternary; the principal rocks are

10 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano basalts, dacites, rhyodacites, and andesitic The area of the Poopó Lake watershed is lavas (Fig. 5). 24,013 km2 and the basin includes 22 Lacustrine sediments originated from vast ephemeral river sub-basins (Calizaya, 2009) inland seas, i.e., palaeolakes that during long (Fig. 1). The Poopó and Uru-Uru lakes are periods covered much of the Altiplano from shallow, with maximum depths of 3.5 and the Pleistocene epoch until around 10,000 0.8 m, respectively (Pillco and Bengts- B.P. (Fig. 5; Fig 20a) (Servant-Vildary et al., son, 2006). Most groundwater recharge in 1993). The withdrawal of the paleo-lakes left the study area occurs during the wet season behind large amounts of salt in the soils and (Garreaud and Aceituno, 2000); the western water bodies. limestone and sandstone formations are important recharge zones due to their high

Figure 4. Affected areas: a) River highly polluted; b) Sludge from a mineral processing plant being discharged into a waste dam; Unaffected areas: c) A shepherd driving a flock of llamas; d) Quinoa plantations; e) River non-affected.

11 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03 permeability (Van Damme, 2002) and cons- basin (Paper I & II). In the second stage, to titute potentially productive aquifers (Fig. 5). characterize the geochemistry of As and Extensive precipitation in the eastern Cordi- other TEs and to assess the extent of llera during the rainy season plays a critical geogenic As distribution in surface and role in the local hydrologic processes groundwater, one set of samples was collec- through surface runoff through the plains. ted from nineteen drinking water wells and The general flow pattern in the plain is three rivers in the southern part of the towards the Lake Poopó. The groundwater Poopó Lake basin (Paper III). Another set flow is slow due to low hydraulic gradient of samples was collected from three rivers, and low permeability of the sediments. forty two drinking water wells and twelve Geothermal springs are commonly encoun- installed piezometers to investigate the quali- tered in the BA region and occur in three ty and the physicochemical characteristics of main areas: Western Cordillera, the Altiplano drinking water and to assess the extent of border of the Eastern Cordillera and the geogenic As distribution in surface and Eastern Cordillera (Huaranca Olivera and groundwater (Paper IV). Water samples Neumann Redling, 2000). The frequent from the same sites were collected in dif- occurrence of geothermal springs near the ferent seasons for monitoring the temporal border of Eastern Cordillera and the Alti- variations in surface and groundwater com- plano corresponds to the face of a very deep positions (Paper III & IV). fault line. Product of a normal heat flow at Prior to sampling at each site, the physico- 2,000 to 3,000 m, these hot waters take chemical parameters such as temperature, advantage of this fracture zone to rise up to pH, Eh, and electrical conductivity (EC) the surface. The origin of the thermal waters were measured in the field, using a portable in the Eastern Cordillera is related to the meter equipped with a combination electro- complex geological setting in the Andes de (Hanna Instruments 9828 multipara- mountains. The Andes have been subjected meter). Sensors and probes in the portable to multiple compressive tectonism, and ac- meter were calibrated each day before sam- companied phases of plutonic magmatism pling using standard calibration solutions. during the Paleozoic and a second phase of Field measured Eh values were corrected for sub-volcanic to extrusive magmatism during the standard hydrogen electrode. The alkali- - Late Miocene period. The geothermal nity (HCO3 ) was also determined in the field springs are mainly associated with folded by titration of a sample aliquot with HCl Paleozoic strata, linked to local faults and until the equivalent point (HACH, Digital fractures. There is no evidence of a direct titrator). For geothermal water, alkalinity was relationship of geothermal waters to igneous measured only after cooling (approx. 5- bodies at the surface; although such 10 min) to 35ºC. During the on-site investi- relationship are known to exist in form of gations, each sampling site was georeferen- hot magma chambers as heat sources at ced using a hand-held global position system depths of 3,000 m (Huaranca Olivera and (Garmin GPSmap60Csx) (Paper I, II, III & Neumann Redling, 2000). IV). 4. MATERIALS AND METHODS Water sampling from each site involved collection of: 4.1. Field investigations i) Filtered samples (Sartorius 0.45 µm filter) 4.1.1. Surface and groundwater sampling for analyses of major anions (Paper I, II, In the first stage, water samples from four III & IV). rivers, twenty-eight wells and sixteen hot ii) Filtered samples (Sartorius 0.45 µm filter) springs were collected to investigate the and acidified (ultrapure HNO3 1%v/v) extent of As distribution and to assess the for major cation and TEs analyses. (Paper hydrogeochemical characteristics of surface I, II, III & IV). and groundwater within the Poopó Lake

12 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Figure 5. Map of the Poopó Lake basin with major geological characteristics and main villages.

13 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

iii) Field speciation of As(III), where effects on the surface and groundwater water sample was filtered through chemical composition (Paper III). Disposable Cartridge® and later 4.2. Laboratory analysis acidified, following procedure outlined by Meng et al. (2001) for 4.2.1. Water analysis - - 2- As(III) analyses (Paper II & III). All Major anions (Cl , NO3 and SO4 ) were samples were stored at low tempera- analyzed at the laboratory of the Division of ture (4°C) until further analyses. Land and Water Resources Engineering at 4.1.2. Sediment sampling KTH using a Dionex DX-120 Ion- Chromatograph (IC). The major cations A total of twelve sediment samples from (Na+, K+, Ca2+, Mg2+), As and TEs were geothermal springs were collected for analy- analyzed in the filtered acidified water sis of As and other TEs; sediments samples samples using a Varian Vista Pro Inductively were taken at the same sites than geothermal Coupled Plasma Optical Emission Spectro- water samples in most places; however, in metry (ICP-OES) at Stockholm University some sites access was not feasible (Paper II). (Paper I, II & III). 4.1.3. Core-sediment sampling For Paper IV, TEs including As were Based on the spatial distribution of dissolved determined by inductively coupled plasma- As in shallow aquifer and because the excep- mass spectrometry (Agilent 7500) at the tionally high As concentrations measured in laboratories of the Scientific Instrumentation wells in the area, core-sediments samples Centre of the University of Jaen in Spain. were collected from the unconsolidated

Quaternary sediments at two different sites; 4.2.2. Geochemical characterization of sed- Condo K and 3 km west of Quillacas, both iments from geothermal springs sites located in the southern part of Poopó Sediment samples from geothermal springs Lake basin (Fig. 5). were dried in an oven at 35°C and sieved (< 2mm) to homogenize the size of the sam- An auger drill was used to collect the ple. A mass of 0.5 g was digested with sediments from representative depths up to HNO (bi-distilled acid, 96.2%) using 3000 7.5 m at Condo K and up to 5 m near 3 Anton Paar microwave digestion system at Quillacas. Continuous sediment cores were 280°C at maximum pressure of 8600 kPa. collected during drilling at each site; eight After digestion, samples were filtered thro- sediment samples were obtained at each ugh 0.45 µm filter and diluted to 50 mL with profile; later, core-samples were split for deionized water solutions were analyzed for lithological studies and sequential extrac- elemental composition including As, using a tion. Different layers were characterized by Varian Vista Pro Inductively Coupled Plas- color, grain size and texture (Paper III). ma Optical Emission Spectrometry (ICP- 4.1.4. Rock sampling OES) at Stockholm University (Paper II). With the help of a geological map which 4.2.3. Geochemical characterization of core display the arrangement of main geological sediments features including different rock types and Sediment samples collected from two auger eras, a total of fourteen rock samples were drills sites were analyzed for grain size, collected in the southern part of Poopó Lake differentiating as, coarse sand, fine sand, basin, comprising a large diversity of rock medium sand, silt and clay. A total of 8 types represented by stratovolcanic, sedi- representative samples were selected at each mentary and rhyolitic volcanic rocks from profile to identify the potential As host different geological eras in order to inves- phases in the aquifer sediment by a 5 step tigate their chemical composition, natural sequential extraction procedure following chemical background, the relation between methods described by Guo et al. (1997), the local geologic setting and especially their Bhattacharya et al. (2006) and Claesson et al. (2003). Details on the sequential extraction

14 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano procedure are given in Supplementary 4.4.2. Rock and sediment samples Materials to Paper III. Each extract from Reagent blanks and spiked samples were sequential extraction was preserved by acidi- prepared and analyzed for quality control of fication with 0.5 mL 14 M HNO3/100mL. TEs in sequential extraction and HNO3 The extracts were later analyzed using extraction in sediments and rock samples. Varian Vista Pro Inductively Coupled Plas- Replicate analyses of samples were carried ma Optical Emission Spectrometry (ICP- out to check the precision of the results; OES) at Stockholm University (Paper III). accuracy and precision indicate variations 4.2.4. Rock analyses within the range of ± 13%. Rock samples were grinded (< 0.125 mm) at 5. RESULTS Åbo University. A mass of 2 g was digested 5.1. Hydrochemistry of groundwater into conical flasks with 15 mL of 7 M HNO3 on a sand bed for 2 h. After digestion, and surface water within the Poopó samples were filtered through 00K filter and Lake basin diluted to 50 mL with de-ionized water; 5.1.1. Groundwater solutions were analyzed for elemental com- Twenty-eight wells were selected for position including As, using a Varian Vista groundwater sampling around the Poopó Pro Inductively Coupled Plasma Optical Lake in the BA. It was evident that only one Emission Spectrometry (ICP-OES) at of the investigated wells was affected by Stockholm University (Paper III). mining, while the others deliver natural 4.3. Geochemical modelling and groundwater from the shallow aquifer of the data analyses Poopó Lake basin. Details of these findings To evaluate the water chemistry results, have been presented in Paper I. Twenty- Aquachem software (4.0.264 Waterloo Hy- seven drinking water samples from shallow drogeologic Inc., 2003) was used to create dug-wells (< 10 m) revealed pH values Piper diagrams, Box & Whisker diagrams, ranging from 6.4 to 7.9 with an average of Giggenbach and bivariate plots Aquachem 7.2, while Eh indicated slightly reduced was also used to determine water types and conditions in some locations, but oxidized in to calculate the ion-balance of water samples others; Eh values ranged between +141 and (Paper I, II, III & IV). Saturation indices and +218 mV with an average of +179 mV. As speciation were calculated using the Electrical conductivity varied in a wide range PHREEQC program (Parkhurst and from 210 to 4,630 µS/cm (average: Appelo, 1999) using the MINTEQ database 1,200 µS/cm). (minteq.dat). Thermodynamic data for Ca- Regarding major cations, the concentrations + arsenate minerals and complexes (Bothe and of Na ranged between 14.6 and 1,333 mg/L Brown, 1999) and Mg complexes with As with an average of 165.7 mg/L; Ca2+ bet- (Whiting, 1992) were added to identify ween 10.0 and 234.5 mg/L with an average additional aqueous species and saturation of 72.6 mg/L, Mg+2 between 3.4 and + indices (Paper I, II & III). 85.5 mg/L (average: 21.6 mg/L) and K between 3.0 and 42.1 mg/L (average: 4.4. Quality control 15.1 mg/L). As regards to anion concen- - 4.4.1. Water samples trations, HCO3 ranged between 36.6 and - For quality control, reagent blanks and 717.6 mg/L with an average of 273.0, Cl between 7.1 and 1,220 mg/L averaging spiked samples were prepared and analyzed. 2- Accuracy was tested by analyzing certified 153.7 mg/L, SO4 between 9.1 and 562.5 mg/l with an average of 121.8 mg/l, reference material NIST SRM 1643 e – TEs - in water. Replicate analyses of samples were and NO3 from below detection level carried out to check the precision of the (0.2 mg/L) and 234.7 mg/L with an average results; accuracy and precision indicated of 21.2 mg/L, indicating an anthropogenic variations within the range of ± 10%. contamination in drinking water at some

15 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Figure 6. a) Piper diagram of groundwater; b) Piper diagram of Antequera River. locations such as Quillacas and Totoral. The centrations of other TEs i.e., Al, Cu, Cd, Fe, most common water compositions are Na- Mn, Si and Zn were very high, showing a

Ca-HCO3 type (12.5%) and Mg-Ca-Na- typical composition of water impacted by HCO3-SO4 type (12.5%) with a geochemical mining (Paper I). evolution trend towards Na-Cl type (Fig. 6a) 5.1.2. Surface water (Paper I). Four sampling sites along Antequera River Among redox sensitive elements, dissolved were considered for the physicochemical As concentrations in drinking water wells characterization and assessment of mining indicate a wide range between below detec- impacts on surface water quality in a mining tion level (5.6 µg/L) and 242.4 µg/L with an region located to the eastern part of Poopó average of 73.4 µg/L (n=27). Likewise, Fe Lake basin. A river water sample from and Mn concentrations also show a wide Chapana, the upper sampling site, located in range of variability, 8.8-1,892 µg/L (average: the headwater area had a pH of 7.6 and EC 112 µg/L) and 0.6-1,089 µg/L (average: of 280 µS/cm. As river flows downstream, 74 µg/L), respectively. Other TEs occur in surface water quality is affected by mining significant concentrations; Al ranged from operations and the huge amounts of mining 13.9 to 403 µg/L with an average of waste deposited in the river banks, thus 48.9 µg/L, Zn ranged from 6.9 to 214 µg/L decreasing pH to 3.9 and to 3.1, and (average: 34.9 µg/L), Si from 7,245 to increasing EC up to 2,580 µS/cm and to 21,717 µg/L (average: 14,337 µg/L), B from 2,530 µS/cm in Totoral and Avicaya, 307 to 6,071 µg/L (average: 1,494 µg/L) and respectively. Finally, reaching the lower sam- Li from 41.9 to 4,256 µg/L (average: pling site at the community of Pazña, the 584 µg/L). Dissolved Cd and Pb had very river gets affected by geothermal water, low concentrations being far below the causing an increase of pH to 4.7 and a WHO guidelines. The only water well affect- decrease of EC to 1,960 µS/cm. Findings ted by mining is located to the eastern part related to major ions concentrations have of the Poopó Lake basin, in a region with been described in detail in Paper I; at high mining activity; this water sample Chapana, the headwater, is of Ca-Mg-Na- shows a pH value of 3.79, Eh of +378 mV HCO -SO type with a geochemical evolu- and EC of 4,500 µS/cm and a chemical 3 4 tion of a Ca-SO4 type in the middle part of composition of Al-Ca-Mg-SO4 type. Regar- the river and finally to a Ca-Na-SO4 type in ding TEs, dissolved As concentration was the lower part of the river (Fig. 6b). relatively low (5.9 µg As/L); however, con-

16 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

In Chapana, the upper sampling site, dissol- 105 mg/L), Ca2+ between 9.7 and 200 mg/L ved As concentration was of 32.4 µg/L, (average: 77 mg/L) and Mg2+ between 2.4 while in the middle part of the river, concen- and 196 mg/L (average: 31 mg/L). Regar- tration decreased to 13.1 µg/L in Totoral ding major anions, Cl- ranged between 177 and below detection levels (<5.6 µg/L) in and 21,022 mg/L (average: 3,781 mg/L), - Avicaya and Pazña. In contrast, Fe concen- HCO3 between 429 and 1,892 mg/L 2- trations increased sharply, from 20.6 µg/L in (average: 997 mg/L), SO4 between 8.0 and - Chapana, to 35,229 µg/L in Totoral, while in 746 mg/l (average: 119 mg/L) and NO3 Avicaya and Pazña it decreased to from below detection level (0.2 mg/L) to 2,817 µg/L and 546 µg/L, respectively. 55 mg/L (average: 5 mg/L). Chemical com- Similar behavior can be observed for other positions of geothermal waters are mainly of heavy metals and TEs such as Mn, Zn, Al, Na-Cl type (37.5%) and of Na-Cl-HCO3 Cd, Pb, and Si (Paper I). type (37.5%) (Fig. 7a). Regarding TEs, 5.1.3. Geothermal springs dissolved As concentrations in geothermal springs were relatively low compared to Water samples collected from sixteen geo- those in cold groundwaters (Paper I & II); thermal springs (Paper II), evidenced pH concentrations vary from 7.8 to 65.3 µg/L values varying from 6.3 to 8.3 (average: 7.0) (average: 23.2 µg/L); As speciation indicates while Eh showed slightly reduced environ- the predominance of As(III) over As(V) spe- ments with values varying from +106 to cies in 65% of samples. Other TEs occur in +204 mV (average: +172 mV). Measured high concentrations and wide ranges, being EC showed high values ranging from 1,860 most relevant elements, Si between 8,336 to 75,810 µS/cm (average: 13,643 µS/cm) and 13,570 µg/L (average: 10,600 µg/L), Li and temperature was in the range of 40 to ranging from below detection level 75°C (average: 56°C). Ranges of concentra- (1.8 µg/L) to 15,439 µg/L, Fe between 22.0 tions of major ions were significantly higher and 826 µg/L with an average of 273 µg/L, than those found in the cold groundwaters; + Rb between 150 and 6,092 µg/L (average: Na ranged from 339 to 12,886 mg/L with 2,151 µg/L) and Sr from 650 to 4,492 µg/L an average of 2,635 mg/L, while concentra- (average: 1,963 µg/L ) (Fig. 7b), other TEs tions of K+, Ca2+ and Mg2+ were generally + were present in relatively lower concentra- low compared to Na . Concentrations of tions (Paper II). K+ ranged from 14.2 to 328 mg/L (average:

Figure 7. a) Piper diagram of thermal springs; b) Selected TEs in thermal springs

17 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

5.2. Hydrochemistry of groundwater Concentrations ranges of major ions were and surface water in the southern part very wide; concerning cations, Na+ concen- of Poopó Lake basin trations ranged between 15.8 and 735 mg/L (average: 172.1 mg/L), Ca2+ between 8.7 and 5.2.1. Groundwater 179.3 mg/L (average: 46.1 mg/L) while In an initial stage only nineteen groundwater Mg+2 and K+ were present in ranges from samples were collected from shallow dug 3.7 to 120.8 mg/L (average: 15.4 mg/L) and wells (Paper III); however, in a second stage 3.2 to 24.8 mg/L (average: 202 mg/L), res- the number of well water samples increased pectively. Concerning concentrations of ma- - up to forty-two, including those of the initial jor anions, HCO3 ranged between 34.2 and stage, in addition, twelve groundwater sam- 1,025 mg/L (average: 265.4 mg/L), Cl- ples were collected from newly installed between 12.7 and 1,109 mg/L (average: 2- piezometers, making a total of fifty-four 224 mg/L), SO4 between 17.7 and - groundwater samples from shallow aquifer 470.2 mg/l (average: 90.5 mg/L) and NO3 of the southern part of the Poopó Lake from below detection levels (0.2 mg/L) to basin (Paper IV). For monitoring the variati- 420.5 mg/L (average: 29.5 mg/L), the last ons in the physicochemical characteristics indicating an anthropogenic contamination and As concentrations, the drinking water in drinking water at some locations such as wells and the piezometers were sampled in Guadalupe, Pampa Aullagas and Quillacas two different seasons (Paper IV). (Fig. 5) (Paper IV). The principal water types

Water samples collected from shallow dug found are Na-Cl-HCO3 type (38.1%) and wells (1-9 m depth; average 3 m; n=42; dry Ca-Na-Cl-HCO3 type (19.1%) with a geo- season) revealed pH values corresponding to chemical evolution trend towards Na-Cl type slightly alkaline conditions with values ran- (Fig. 8a). ging from 5.5 to 8.7 with an average of 7.8, Concerning TEs, concentration of dissolved while Eh values ranged from +70 to As in drinking water wells ranged from 3.3 +461 mV (average: +294 mV), indicating to 433.4 µg/L (average: 99.2 µg/L) (Fig 9b), moderately reduced conditions in some pla- which is a major concern for the exposed ces, but oxidized conditions in others. The population (Paper IV). Arsenic speciation in- EC varies in a wide range from 295 to dicates the predominance of As(V) over 4,373 µS/cm (average: 1,236 µS/cm) and As(III) species, which is coherent with pH DO varies from 0 to 3.4 mg/L (average: and Eh values found in the groundwater 1.2 mg/L). samples (Paper III). Distribution of other

Figure 8. a) Piper diagram of groundwater; b) Selected TEs in drinking water wells.

18 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

TEs in drinking water wells are also widely between 34.1 and 6,606 mg/L (average: - variable (Fig. 8b). Aluminum concentrations 1,475 mg/L), HCO3 between 133.7 and 2- range from 0.6 to 53.5 µg/L (average 1,434 mg/L (average: 355.8), SO4 between 4.3 µg/L), Li from 46 to 3,594 µg/L 21.2 and 1,714 mg/L (average: 412.3 mg/L). - (average 850 µg/L), Zn concentrations range Concentrations of NO3 were in the range of from 0.8 to 3,068 µg/L (average: below detection level (0.2 mg/L) and 117.7 µg/L), and Sr ranging from 61.8 to 8.0 mg/L (average: 1.0 mg/L). The predo- + - - 3,353 µg/L (average: 568.3 µg/L). Among minant ions were Na , Cl and HCO3 , and redox sensitive elements, Fe and Mn show hence the most common water composi- wide variability in the ranges of 4.2- tions correspond to Na-Cl type (33%) and

484.4 µg/L (average 115.1 µg/L) and 0.3- Na-Cl-HCO3 type (25%) with a geochemical 3,984 µg/L (average 192.4 µg/L), respective- evolution trend towards Na-Cl type (Fig. 8a). ly (Paper IV). Boron concentrations also Concerning the presence of TEs, the con- exceed the WHO guideline (500 µg/L) centration of dissolved As ranged from 12.7 (WHO, 2011) in drinking water wells to 3,497 µg/L (average: 404.8 µg/L), being ranging between 507 and 4,359 µg/L 3,497 µg As/L so far the highest concentra- (average: 1,902 µg/L; n=15) (Paper III). tion of geogenic As recorded in groundwater Twelve piezometers recently installed in the of the BA. Other TEs were also present in Quaternary unconsolidated sediments, bet- wide ranges of concentrations, Fe and Mn ween Condo K and Quillacas (Fig. 5) were concentrations ranged between 25.2 and monitored for physicochemical characteris- 915 mg/L (average: 281.1 mg/L) and 29.5 to tics of groundwater and for chemical com- 4,516 mg/L (average: 747.5 mg/L), respecti- position variation during different seasons vely. Aluminum concentrations ranged from (Paper IV). Piezometers depths were in a 0.5 to 9.6 µg/L (average: 3.5 µg/L), Li bet- range from 2 to 5.3 m (average: 3.8 m) while ween 106 and 31,570 µg/L (average: groundwater levels were in a range of 0.9 to 7,150 µg/L), Zn between 0.6 and 37.6 µg/L 1.5 m (average: 1.2 m). Water extracted from (average: 8.6 µg/L) and Sr from 224.3 to the piezometers is representative of the 10,488 µg/L (average: 2,292 µg/L). shallow aquifer in the southern part of the 5.2.2. Surface water Poopó Lake basin. Groundwater had similar characteristics of wells-water samples; pH Rivers are used in some areas for irrigation values indicate slightly alkaline conditions although in some locations also for human with pH values ranging from 7.2 to 8.6 consumption. In an initial stage, rivers were (average: 7.8), while Eh ranged from +18 to dry and only one water sample was obtained +376 mV (average: +277 mV). Dissolved (Paper III); however, in the second stage, oxygen was not detected; probably oxygen three major rivers were sampled, collecting might not be present because piezometers three samples along the Marques River, two were newly installed. Electrical conductivity samples along the Sevaruyo River and one also showed a wide variability although sample on Cortadera River. Sites along the much higher than found in the water of three rivers were sampled in two different water-wells, ranging from 368 to seasons for monitoring seasonal variations 20,900 µS/cm (average: 4,892 µS/cm). on their physicochemical characteristics and variations of As concentrations (Paper IV). The major ion concentrations were in differ- + Main physicochemical characteristics shown rent ranges. The concentrations of Na ran- pH values corresponding to slightly alkaline ged between 30.8 and 7,240 mg/L (average: conditions with values ranging from 8.2 to 1,622 mg/L), Ca2+ between 14.5 and 2+ 8.7 (average: 8.5) while Eh ranged from +4.3 419.8 mg/L (average: 92.6 mg/L), Mg to +284.4 mV (average: +155.8 mV). Elec- between 2.4 and 179 mg/L (average: + trical conductivity was lower than those 41.7 mg/L) and K between 8.2 and measured in groundwater, ranging from 452 215.4 mg/L (average of 52.5 mg/L) while - to 2,076 µS/cm (average: 872 µS/cm). for major anion concentrations, Cl ranged Dissolved oxygen in surface water was

19 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Figure 9. a) Piper diagram of surface water; b) Selected TEs in rivers. particularly higher compared with ground- (average 34.9 µg/L) and 0.8-3.4 µg/L water, DO in rivers occur in the range of (average 2.4 µg/L), respectively. Other TEs 2.5-6.6 mg/L (average: 4.0 mg/L). Tempe- occur in much lower concentrations than rature varied considerably due to environ- those found in groundwater (Fig. 9b). mental settings, ranging from 8.1 to 24.9°C Aluminum concentrations were in the range (average: 14.4°C). of 1.6-21.9 µg/L (average: 8.7 µg/L), Li Major ions in surface water occur in lower ranged between 1,060 and 4,359 µg/L concentrations than those found in ground- (average: 1,907 µg/L), Zn ranged from 0.8 water; Na+ concentrations ranged between to 6.3 µg/L (average: 2.4 µg/L), and Sr 60 and 336 mg/L (average: 20.4 mg/L), Ca2+ concentrations were in the range of 191.9- ranged from 6.8 to 24.6 mg/L (average: 729.8 µg/L (average: 367.3 µg/L) 16 mg/L), Mg2+ between 2.6 and 13 mg/L (Paper IV). + (average: 7.2 mg/L) and K between 9.3 and 5.3. Geochemical modelling 43.9 (average 20.4); regarding major anions Results of speciation modeling in ground- Cl- concentrations ranged from 95.2 to - water within the Poopó Lake basin indicate 496.7 mg/L (average: 197.2 mg/L), HCO 3 As(V) as the predominant species; main from 95.2 to 457 mg/L (average: 2- 2- species are distributed as HAsO (56.3%), 220.3 mg/L), SO between 15.7 and 4 4 H AsO - (17.4%), CaHAsO (15.2%) and 56.8 mg/L (average: 32.7 mg/L) and NO - 2 4 4 3 MgHAsO (9.1%) (Paper I). from between below detection level 4 (0.2 mg/L) to 2.2 mg/L (average: 0.7 mg/L). Regarding geothermal springs, speciation Piper diagram shows a clear chemical com- modeling indicates that As(III) species is position of surface water being Na-Cl-HCO predominant over As(V) species distributed 3 0 - type the main chemical characteristic in the as of H3AsO3 (61.1%), H2AsO3 (3.9%), 2- - studied rivers with a geochemical evolution HAsO4 (22.1%) and H2AsO4 (8%) trend towards Na-Cl type (Fig. 9a). (Paper II). Regarding TEs, concentration of dissolved Results of speciation modeling in ground- As in rivers ranged from 8.6 to 117.4 µg/L water of fifteen drinking water wells within (average: 73.6 µg/L) (Fig. 9b); almost all the southern part of the Poopó Lake basin, river water samples contain elevated concen- revealed As(V) species predominance over 2- trations of As by far exceeding the WHO As(III) species, distributed as HAsO4 - guideline value (10 µg/L) (WHO, 2011). The (51.2%), H2AsO4 (13.4%), CaHAsO4 0 redox sensitive elements, concentrations of (14.1%), MgHAsO4 (4.4%) and H3AsO3 Fe and Mn were in ranges of 10.1-68.8 µg/L (15.2%) (Paper III).

20 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Saturation indices for carbonate minerals, presented in Fig. 10. For detailed description i.e., calcite [CaCO3], dolomite [CaMg(CO3)] of lithological characteristics at both loca- and magnesite [MgCO3], indicate that waters tions the reader is referred to Paper III. are mostly undersaturated or at equilibrium Results of the 7 M HNO3 extraction in the in most samples of groundwaters and ther- sediments from Condo K, revealed AsHNO3 mal springs, but with respect to Fe- concentrations range between 11.1 and oxide/hydroxide minerals i.e., goethite [α- 50.6 mg/kg (average: 24.3 mg/kg) while FeOOH] and ferrihydrite [Fe(OH) ] ground- 3 sediments in Quillacas revealed AsHNO3 waters and geothermal waters are super- concentrations range between 1.6 and saturated (Paper I, II & III). 36.3 mg/kg (average: 10.4 mg/kg). Other 5.4. Geochemical investigations TEs, i.e., FeHNO3 and AlHNO3 in Condo K range between 1.7-12.3 g/kg (average: 5.3 g/kg) 5.4.1. Arsenic and TEs in sediments from and 0.9-3.4 g/kg (average: 1.8 g/kg), respect- geothermal springs tively, with similar concentration levels in

Twelve sediment samples collected from Quillacas, FeHNO3 between 2.6-11.3 g/kg (a- geothermal springs were geochemical cha- verage: 5.3 g/kg) and Al HNO3 between 1.1- racterized to determine TEs concentrations; 3.6 g/kg (average: 1.8 g/kg). Regarding average concentrations of As were relatively CaHNO3, concentrations are higher at Condo high (23.7 mg/kg), while average concentra- K than in Quillacas (average: 17,189 mg/kg tion of other TEs were variable: Fe and 2,697 mg/kg, respectively), probably (28,1 mg/kg), Li (71.7 mg/kg), Mn due to the presence of shells found in (321.7 mg/kg), Si (462.6 mg/kg) and Zn Condo K below 4 m. (63.7 mg/kg) (Paper II). Results of sequential extraction of As from 5.4.2. Arsenic and TEs in core-sediments each sediment layer at both profiles are Sediments samples were collected from two presented in Fig. 11. Arsenic leached during profiles located at Condo K and 3 km DIW extraction is significantly higher in northwest of Quillacas. The successions of Quillacas profile (1.0-8.6 mg/kg) compared the investigated sediments are schematically to Condo K profile (1.1-4.1 mg/kg). In comparison with the other extraction steps,

Figure 10. Lithological successions of sediment samples collected from a) Condo K and b) Quillacas.

21 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Figure 11. Sequential extractions of As from each sediment layer in a) Condo K and b) Quillacas. relatively small amounts of As (undetected (average: 107 mg/kg). For geographical loca- to 2.8 mg/kg) were released during HCO3 tions of the sampling sites, the reader is refe- extraction from both analyzed profiles. The rred to Paper III. acetate extraction released high amounts of ISCUSSION As; 2.6-11.0 mg/kg in Condo K and 1.3- 6. D 6.6 mg/kg in Quillacas. Oxalate extractable 6.1. Origin of surface and ground- As concentration is significantly higher in water composition the sediments from Condo K (range: 1.4- Sodium and Cl- in groundwater around Poo- 23.1 mg/kg compared to Quillacas (range: pó Lake show a strong positive correlation bdl-6.5 mg/kg). For other TEs concentra- (r=0.897) (Fig. 12a), similar behavior was tions in both profiles, the reader is referred found in the shallow aquifer of the southern to Paper III. part of the Poopó Lake basin (r=0.862) 5.4.3. Arsenic and TEs in rocks (Fig. 12c) suggesting that groundwater is Concentrations of As in rock samples ran- influenced by dissolution of evaporites i.e. ged between 0.3 to 27.1 mg/kg with an halite, which occurs in large amounts as a average of 5.7 mg/kg (Paper III). Volcanic consequence of the arid conditions and high rock from Los Frailes formation showed low rate of evaporation within the study area As concentration between 0.9 to 0.5 mg/kg (Paper I & IV). in Cayco Bolivar and Estación Marques, Water of geothermal springs also show respectively, while the highest concentra- strong positive correlation between Na+ and tions of As were found in coral limestone Cl- (r=0.999) (Paper II) but in much higher rock close to Bengal Vinto village with As concentrations than those in groundwaters; concentrations of 27.1 mg/kg. Another rock the high Na+ and Cl- contents in geothermal type with high As concentration were calca- springs are not only attributed to the reous sandstone close to Bengal Vinto and dissolution of halite, but also to enriched evaporites sampled in the road between waters formed through evaporation and Pampa Aullagas and Orioca with concentra- condensation at shallower depths and also tions of 10.5 and 13.1 mg/kg, respectively. probably to leaching of salt from the saline Regarding major element composition, Ca units of the BA (Fig. 13) (López et. al., ranged from 0.1 to 322 g/kg (average: 2012; Morteani et. al., 2014). Halite is not 66.7g/kg), K from 0.3 to 24.2 g/kg (average the only source of sodium in groundwaters 5.1 g/kg) and Fe from below detection level and hot springs; other processes such as to 9.3 g/kg (average: 3.6 mg/kg). Other TEs weathering of plagioclase and ion exchange present in high concentrations were Al within clays and other minerals could influence in the range of 0.1 to 7.0 g/kg (average: the salinity (Risacher and Fritz, 2009; Raych- 2.4 g/kg) and Sr between 4 to 421 mg/kg owdhury et. al., 2014).

22 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Figure 12. Bivariate plots a) Na+ – Cl- in groundwater around Poopó Lake; b) 2+ 2+ 2- - + - Ca +Mg – SO4 +HCO3 in groundwater around Poopó Lake; c) a) Na – Cl in 2+ 2+ 2- - groundwater in the southern part of the basin; d) Ca +Mg – SO4 +HCO3 in groundwater in the southern part of the basin.

Figure 13. Environmental characteristics of thermal springs at: a) Castilluma and b) Vichajlope (Paper II).

23 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

2- - 2+ 2+ The SO4 +HCO3 and Ca +Mg are less of total variability. Results show three sam- correlated r=0.861 (Fig. 12b) and r=0.719 ples that are affected by AMD; two of them (Fig. 12d) for the whole basin and for the are samples strongly affected by mining a- southern part of the basin respectively long Antequera River at Totoral and Avicaya (Paper I & IV), suggesting that main sources locations (TOR1 and AVR1, respectively) of these components are the weathering and the third sample (SORP1) is the only and/or dissolution of gypsum and carbonate water well affected by AMD (Fig. 15, upper minerals, i.e., limestone and dolomite, which right). Two samples are affected by dis- is supported by the saturation indices calcu- solution of salts and AMD (Fig. 15, center); lated for carbonate minerals, i.e., calcite the sample code PAZR1 is the lower

[CaCO3], dolomite [CaMg(CO3)] and sampling site on Avicaya River close to magnesite [MgCO3] that are mostly under- Pazña which is affected by mining and at saturated or at equilibrium in most ground- that location by thermal waters. Two sam- waters and thermal springs suggesting that ples are affected by high salinity caused by groundwater chemistry is controlled merely the dissolution of halite (Fig. 15, lower by the precipitation/dissolution of carbonate right). All the other samples are represen- minerals (Table 3 in Paper I, Fig. 6 in tative of the natural conditions of the Paper II and Table 6 in Paper III). On the shallow aquifer within the Poopó Lake basin other hand, major ion composition of water except for the sample code CHAO1 which from the geothermal springs are affected is the base flow composition at the head- locally by meteoric water or by mixing with water of Antequera River and is comparable shallow groundwater changing their chemi- to groundwater composition although in cal composition from Na-Cl type to a Na- much lower concentrations. Cl-HCO type (Fig. 14) during its ascension 3 6.2. Spatial distribution of As and to ground surface (Paper II). other TEs The principal component analysis (PCA) (Paper I) shows that Principal Component 1 6.2.1. Groundwater within the Poopó Lake 2- - (PC1) has high loadings for EC, SO4 , Cl basin and Na+ and intermediate loadings for Ca2+ Initial screening of geogenic As and other - and HCO3 and explains 86.6% variability in TEs in groundwater was conducted over the date set. Principal Component 2 (PC2) has entire the Poopó Lake basin by sampling 2- 2+ high loadings for SO4 , Ca and Zn and groundwater from twenty seven dug-wells negative loadings for Na+ and Cl- and for the hydrochemical characterization of explains 10.3% of variability in the data set. the shallow aquifer in the area (Paper I). Both PC1 and PC2 explain together 96.9% Arsenic concentrations in groundwater varied spatially across the Poopó Lake basin with no obvious trends. Overall, there were no obvious distribution patterns of As in groundwater except for one area located south of the Poopó Lake, where high concentrations of geogenic As occur in groundwater specially in those wells lying in the Quaternary fluvio-lacustrine formations (Fig. 5). The highest concentrations of As within the Poopó Lake basin were found in two drinking water wells located in Toledo in the northwestern part of the basin (Fig. 16a), Figure 14. Geochemical evolution of with concentrations of 234.3 µg/L and thermal waters through mixing with 242.4 µg/L, respectively. Another four shallow groundwaters. drinking water wells, with high As

24 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Figure 15. Results of Principal Component Analysis. Sample codes inside circles are water affected by AMD and/or by high salinity. Sample codes not circled are groundwater or surface water representative of the natural conditions of the shallow aquifers and the base flow within the Poopó Lake basin, in the Bolivian Altiplano. concentrations were found in Pampa the predominant species distributed as 2- - Aullagas, Quillacas and Condo K, villages HAsO4 (56.3%), H2AsO4 (17.4%), located in the southern part of the Poopó CaHAsO4 (15.2%) and MgHAsO4 (9.1%) Lake basin (Fig. 5, Fig. 16a) where As (Paper I). concentration varied in a range of 116.8 to Fifty two percent of drinking water wells 233.3 µg/L (average: 200 µg/L). Arsenic around Poopó Lake had high B concen- concentrations in other drinking water wells trations in the range of 1,003-5,408 µg/L were randomly distributed from below (average: 2,227 µg/L) (Fig. 16b), and Li was detection level (<5.2 µg/L) to 185.7 µg/L present at high concentrations in four drin- (average: 33.6 µg/L) (Fig. 16a). Speciation king water wells ranging from 1,713 to modeling performed in samples collected 4,226 µg/L (average: 2,559µg/L), in the around Poopó Lake basin, indicates As(V) as remaining drinking water wells Li concentration average was 240 µg/L (Fig. 16c). Other TEs were also widely distributed in drinking water wells, although their concentrations were relatively low and not relevant for drinking quality point of view (Paper I). The spatial distributions of the investigated elements follow a slight similar pattern between B, Li and EC but do not follow any common pattern with As (Fig. 16). The heterogeneous values EC values of groundwater (Fig. 16d) could be explained by the local hydrological characteristics. 6.2.2. Groundwater and surface water in the southern part of Poopó Lake basin After the initial scrutiny on As in shallow aquifers across the entire Poopó Lake basin, and once it was recognized that the more impacted area was located to the south of the basin, a second investigation was con- Figure 16. Spatial distribution pattern of ducted by sampling forty two drinking water a) As, b) B, c) Li and d) EC in drinking wells, twelve piezometers and three rivers in water wells around Poopó Lake. the southern part of the Poopó Lake basin

25 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Figure 17. Spatial distribution pattern of As in groundwater of shallow aquifer in the southern part of the Poopó Lake basin.

(Paper IV). The highest concentration of Quillacas (Fig. 17) and again, another three dissolved As (433.4 µg/L) in drinking water wells at distances below 3 km showed lower was found in a sample collected from a As concentrations within a range of 24.1 to hand-dug well inside a dwelling in Jiska- 105.8 µg/L. Arsenic concentrations in other Collo (Fig. 17); at the same location, another drinking water wells were randomly distri- two wells had lower concentrations (8.3 and buted and similar behavior has been found 11.0 µg/L, respectively). Another drinking for As concentrations in groundwater sam- water well containing high concentration of ples collected from installed piezometers. dissolved As (260.5 µg/L) was found in The area where piezometers are located is of

26 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano approximately 30 km2; the highest concen- stream (56.8 to 80.9 to 85.3 µg/L) (Fig. 17; tration of dissolved As was 3,497 µg/L while Fig. 18); here is assumed that groundwater the lowest was 12.7 µg/L confirming that recharge containing high concentrations of the behavior in the spatial distribution of As As is more relevant than dilution/attenua- is highly variable, even in small areas tion processes; however, to confirm this (Fig. 17). Speciation modeling was perfor- assumption, more detailed hydrological med in fifteen drinking water wells; the investigations must be performed. Only one results indicate As(V) species predominance sample was collected at the lower part of the over As(III) and species are distributed as Cortadera River, just before to discharge 2- - HAsO4 (51.2%), H2AsO4 (13.4%), into Poopó Lake. Water sample from CaHAsO4 (14.1%), MgHAsO4 (4.4%) and Cortadera River shows the highest EC 0 H3AsO3 (15.2%) which is coherent with the (2,076 µS/cm), which is due to the influence measured Eh and pH (Paper III). of a geothermal spring discharge within the Regarding other TEs, only fifteen drinking water course 21 km upstream from the sam- water wells were investigated for the presen- pling point. Arsenic concentrations in this ce of B; in all them, B concentrations excee- river are significant lower compared to those ded the Bolivian regulation and the interna- in the other rivers; during rainy season the tional guidelines (300 µg/L and 500 µg/L, measured concentration was 10.4 µg/L. The respectively); B distribution varies significan- occurrence of high concentrations of dis- tly although the number of wells is not solved As already present in the headwaters enough to observe a pattern (Paper III). show how easily it can be mobilized from Boron concentrations ranged between 507 the volcanic fields located in the south- and 4,359 µg/L. Lithium was also present in western part of the study area through the very high concentrations in twelve drinking physical weathering of ignimbrites, sediment water wells in a range of 1,045 to 3,549 µg/L transportation by runoff, and release of As (average: 2,271 µg/L) while in the remaining caused by the sediment-water interactions thirty wells Li ranged from 46 to 904 µg/L which promote As to solution through (average: 282 µg/L) (Paper IV); the spatial dissolution/desorption mechanisms. distributions of Li do not follow any com- 6.3. Seasonal variations mon pattern with As. Other TEs in ground- Regarding seasonal variations, with few water were present in low concentrations exceptions, As concentrations in ground- and not relevant for drinking quality point of water trend to decrease during dry season; view (Paper IV). Weak positive correlations however, the change is not very significant were observed in groundwater between As (Paper IV). Major ions decrease in same pat- and pH (r=0.286), As and EC (r=0.430), As - + tern as As, with the consequent diminution and Cl (r=0.433), As and Na (r=0.514), As - 2+ of EC (Table 2). Other TEs such as Fe, Mn, and HCO3 (r=0.359), As and Ca 2- Li and Zn also decrease during dry season. (r=0.375), As and SO4 (r=0.498), and As and Sr (r=0.409), and does not correlate well Seasonal variation indicate similar pattern with other parameters (r<0.15) (Paper IV). along rivers in both rainy and dry seasons (Paper IV). Arsenic concentrations are lower Regarding surface waters, concentrations of during rainy season due to dilution while dissolved As along rivers do not show signi- during dry season, As concentrations are ficant variations; water samples collected higher due to the high rate of evaporation, during rainy season along Sevaruyo River very common characteristic in the semi-arid indicate from the upper sampling point, a zone of the BA (Fig. 18). Most major ion decreasing trend in As concentrations concentrations increase downstream in downstream (117.4 to 92.5 µg/L) (Fig. 17; Sevaruyo and Marques rivers with the Fig. 18), which is probably due to natural consequent increasing of EC (Table 2), attenuation and/or dilution, while concen- probably due the weathering/dissolution of trations of dissolved As in Marques River, evaporites and mineral phases such as halite, increases from upper sampling point down-

27 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Table 2. Selected statistical parameters of drinking-water wells, rivers and piezometers at different seasons. Rainy season (April 2013) Dry season (September 2013) Parameter Water source min. max. average median min. max. average median Wells 6.1 8.9 8.2 8.3 5.5 8.7 7.8 7.9 pH Rivers 8.8 9.6 9.1 9 8.2 8.7 8.5 8.5 Piezometers 7.5 8.9 8.3 8.3 7.2 8.6 7.8 7.9 Wells 318 4,743 1,349 900 295 4,373 1,232 832 EC Rivers 316 1714 769 652 452 2,076 872 539 (μS/cm) Piezometers 389 21,600 5,125 1,537 368 20,900 4,892 1,280 Wells -87 567 162 139 70 461 294 295 Eh Rivers 100 433 188 156 214 494 366 339 (mV) Piezometers 17 698 164 124 18 376 247 277 Wells 0 2.9 0.8 0.7 0 3.4 1.2 1.1 DO Rivers 3.4 4.8 4.1 4 2.5 6.6 4.4 3.3 (mg/L) Piezometers nd nd - - nd nd - - Wells 65.9 1,413 340.7 278.2 34.2 1,025 265.4 217.5 HCO - 3 Rivers 102.5 453.8 222 186.7 95.2 457 220.3 199.8 (mg/L) Piezometers 175.7 1,735 449.6 292.8 133.7 1,434 355.8 189.8 Wells 15 1,127 259.5 141 12.7 1,109 224 120 Cl- Rivers 54.6 394 177.4 150.7 95.2 496.7 197.2 146.5 (mg/L) Piezometers 34.8 6,697 1,557 436.2 34.1 6,606 1,475 269.7 Wells <1 677.1 128 86.7 17.7 470.2 90.5 53.2 SO 2- 4 Rivers 12.7 58.4 32.5 27.8 15.7 56.8 32.7 30.2 (mg/L) Piezometers 33 2,343 460.5 61.8 21.2 1,714 412.3 39.5 Wells 23.1 1,180 227.5 142.2 15.8 735 172.1 89.8 Na+ Rivers 49.4 320 143.7 141.5 60 336 147.6 117.9 (mg/L) Piezometers 40.2 7,740 1521 214.3 30.8 7,240 1,622 220.3 Wells <0.1 244 29.7 14.8 3.2 202 24.8 16.3 K+ Rivers 4.9 29 12.6 10.4 9.3 43.9 20.4 15.7 (mg/L) Piezometers 8.8 190 45.8 14.7 8.2 215.4 52.5 17.4 Wells 1.7 140.1 44.5 32.8 8.7 179.3 46.1 30.5 Ca2+ Rivers 6 20.6 12.4 12 6.8 24.6 16 16 (mg/L) Piezometers 14 384.6 103.4 72.4 14.5 419.8 92.6 64.6 Wells 3.7 59.4 13.7 9 3.7 120.8 15.4 8.2 Mg2+ Rivers 2.2 8.7 5.2 4.9 2.6 13 7.2 7 (mg/L) Piezometers 2.7 193.8 44.8 18.7 2.4 179 41.7 15 Wells 3.5 622.9 118.7 79.8 3.3 433.4 99.2 73.6 As (µg/L) Rivers 10.4 91.4 62 69.2 8.6 117.4 73.6 83.1 Piezometers 22.9 2,475 376.8 219 12.7 3,497 404.8 138.7 Wells 20.3 472.9 117.3 78.4 4.2 484.4 115.1 75.4 Fe Rivers 8.3 55.1 27.3 24.9 10.1 68.8 34.9 26.5 (µg/L) Piezometers 57.1 914 326.8 197.4 25.2 915 281.1 190.1 Wells 0.9 2,787 154.5 6.6 0.3 3,984 195.4 2.1 Mn Rivers 1.3 15.3 3.9 1.7 0.8 3.4 2.4 2.5 (µg/L) Piezometers 72.3 4,619 958.6 353.4 29.5 4,516 747.5 151.2 Wells 52.6 3,969 922.9 391.1 46 3,594 850 339.5 Li Rivers 626.4 3,431 1,397 1,121 1,060 4,359 1,907 1,489 (µg/L) Piezometers 118.6 27,338 7,419 2,697 106 31,570 7,150 1,267 Wells 15.6 4,034 153.8 24.9 0.8 3,068 117.7 6.6 Zn Rivers 22.6 52.5 30.3 26.3 2.7 6.5 4.7 4.7 (µg/L) Piezometers 16.3 34.5 23.4 22.2 0.6 37.6 8.6 2.8 nd: not detected; -: not calculated.

28 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

concentrations in geothermal springs exceed 1 mg/L (Ball et al., 1998) and over 100,000 kg of geothermal As is estimated to leave the western boundary of the Park each year (Nimick et al., 1998). In Chile, the concentration of As at the geothermal field El Tatio reaches 27 mg/L and concentration of about 1.4 mg/L were detected downstream in the Rio Loa fed by the discharge of geothermal water (Romero et al., 2003). In the BA, Banks et al. (2004) found extremely high As concentrations draining from areas of geothermal springs and fumaroles in the Western Cordillera, concentrations in the surface waters reached up to 4.6 mg/L (median 34 µg/l). For this thesis, the initial Figure 18. Seasonal variation of As hypothesis led to the study geothermal concentrations in surface water. Filled springs as potential sources for groundwater symbols correspond to the dry season and surface water contamination resulting in while open symbols correspond to high concentrations of As. However, dis- rainy season. Triangles correspond to solved As concentrations in geothermal Sevaruyo River, diamonds correspond waters assessed during the present thesis to Marques River and circles corres- were very low compared to those reported pond to Cortadera River. by Banks et al. (2004), and besides relatively low compared to As concentration in limestone and calcite. In Cortadera River, groundwaters of the shallow aquifer in the seasonal variation of As concentrations have Poopó Lake basin (Paper I & II). Discharges opposite behavior compared with other of geothermal waters only reach up to 65.3 rivers, probably due to the effect of µg/L (average: 23.2 µg/L) and besides, geothermal spring discharge into the water following the discharge, As is attenuated by course; however, the increment is not several processes including dilution and significant as As concentrations increase adsorption onto stream sediments enriched only from 8.6 µg/L during dry season to in ferric oxyhydroxides (Fig. 13a), this is 10.4 µg/L in the rainy season, (Fig. 18). supported by the saturation indices calcula- Other TEs vary randomly however varia- ted for Fe-oxide/hydroxide minerals i.e., tions are not significant (Table 2) (Paper IV). goethite [α-FeOOH] and ferrihydrite [Fe(OH) ] that are mostly supersaturated in 6.4. Sources of As and main 3 geothermal waters (Paper II); hence it is mechanisms for As mobilization concluded that geothermal water is not a 6.4.1. Geothermal springs significant source for high concentration of At Los Azufres and Los Humeros sites in As in groundwater, at least not for shallow Central Mexico, As concentration in geo- aquifers along the Eastern Cordillera within thermal waters reaches up to 49.6 mg/L and the Poopó Lake basin. 73.6 mg/L, respectively (Birkle et al., 2010; 6.4.2. Rock types López et al., 2012; Bundschuh and Prakash, According to Taylor and McLennan (1985) 2015). Discharge of geothermal waters at the abundance of As in the upper continen- Los Azufres and their leakage from evapora- tal crust is approximately 1.5 mg/kg. Within tion ponds contaminates surface drainage the Poopó Lake basin, most of the indivi- where concentrations of As up to 8 mg/L dual rock types analyzed for As have higher were found (Birkle and Merkel, 2000). In values. Yellowstone National Park (USA), As

29 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Figure 19. a) Map of the BA basin. Modern lakes are in blue and salars are in white. The red line represents the approximate maximum extent of palaeo- Lake Tauca at ~3780 m a.s.l.; b) Outcrop of coral limestone in the hills of Pampa Aullagas; c) Coral limestone wraps the host andesitic rock.

The highest concentration of As was found 10% fine grained, broken, resorbed quartz in coral limestone (27 mg As/kg); the and 5-10% plagioclase (du Bray et al., 1995). average value for limestone/dolomite is 2.6 Weathering/dissolution of limestone and mg As/kg (Baur and Onishi, 1969), hence silicate minerals typically results in an incre- the measured As concentration is ten times ases of pH, alkalinity and base cation higher in the coral limestone within the concentrations in solution; these are favora- study area. Limestone is widely distributed ble for high mobility of oxyanionic As throughout the BA, where withdrawal of species (Banks, 2004; Smedley and Kinni- paleolakes have left and extensive coral lime- burgh, 2002); this is supported by the stones with outcrops visible at many sites, calculated saturation indices for carbonate such as at Pampa Aullagas, Tusqui and minerals, i.e., calcite [CaCO3], dolomite Bengal Vinto (Fig. 19) lying approximately [CaMg(CO3)] and magnesite [MgCO3] that 50 m above the current level of Poopó Lake. are mostly undersaturated or at equilibrium Other rock samples had up to 13.1 mg in most groundwaters of the study area As/kg (evaporite), 10.5 mg As/kg (calcare- (Paper I, II & III) and by the strong ous sandstone), 4.7 mg As/kg (andecitic to correlation (r=0.926) between As and bi- dacitic lava), 1.7 mg As/kg (sandstone) and carbonate found in groundwater samples 1.2 mg As/kg (rhyolitic lava). High concen- collected from piezometers, although corre- trations of other TEs were found in surface lation in drinking water wells were much water and groundwater i.e., Li, B, Sr and Si lower (r=0.359) (Paper IV). within the study area (Paper I, II, III & VI). Many studies indicate that volcanic rock The distribution of these elements in the types are significant sources of As (Nicolli et surface and groundwaters is caused by rapid al., 1989; Smedley and Kinniburgh, 2002; alteration and weathering of volcanic rocks, Banks, et al., 2004). Meteoric water especially the ignimbrites (Risacher and infiltration through the permeable volcanic Fritz, 2009). The major chemical compo- rocks and longer residence time of ground- sition of unaltered volcanic rocks were water with subsequent flushing during rainy reported to have 63.6% of SiO and 16.6% 2 season, causes the release of high amounts of Al O in a lava sample collected from 2 3 of As and other TEs to groundwater of Cerro Gordo (Fig. 17) with 3-5% fine grai- shallow aquifer and surface water in the ned plagioclase and 2-4% hornblende; while headwaters of the Poopó Lake basin. These in a tuff sample collected near Estación factors also show that the Los Frailes Marques (Fig. 17) major composition were volcanic field (Fig. 5) has an extensive 66.0% of SiO and 17.1% of Al O (with 8- 2 2 3 influence on the general water chemistry of the entire study area. However, direct impact

30 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano of the volcanic rocks on the As conta- were 50.6 mg As/kg and 36.3 mg As/kg, mination seems to be relatively limited in respectively. Measured contents of As in the spite of being considered a probable primary sediments were considerably higher than the source for sediments in the region. estimated global average of 1–5 mg/kg 6.4.3. Sediments (Ravenscroft et al., 2009; Tanaka, 1988). The levels were high also compared to other The physical weathering of volcanic and known high-As areas in Argentina, 2.5–7.1 sedimentary rocks produces vast quantities mg/kg (Smedley and Kinniburgh, 2002) and of sediments which due to runoff and wind, Bangladesh 1.0–14.7 mg/kg (Bhattacharya et is redistributed along the highland plain. al., 2006); thus sediments in the area, clearly High dissolved As concentrations in wells play an important role in the As contamina- lying in the unconsolidated formation in the tion. plain point out to the sediments as a major source of As; however, local variations of In the sequential extraction procedure per- the concentration of As in groundwater are formed in the sediments, a strong positive significant and indicate that the sources of correlation is observed between oxalate As are not homogeneously distributed thro- extracted Fe and As in both profiles; ughout the landscape. This is supported by r=0.969 in Condo K and r=0.928 in the lithological succession of sediment Quillacas (Fig. 20), revealing that As is samples collected from Condo K and Quilla- closely associated with amorphous Fe cas (Fig. 10), where the auger drilling oxyhydroxides. Since groundwaters are oxi- revealed a rather high vertical variability in dizing, Fe oxides in the sediments are stable the characteristics of the sediments and also and hence the mobility of As is constrained differences between the two drilling sites by sorption processes; this is supported by (Paper III). The clays, silts and fine sands of the saturation indices calculated for Fe- the lacustrine sediments in Condo K have oxide/hydroxide minerals i.e., goethite [α- more than twice the average As content of FeOOH] and ferrihydrite [Fe(OH)3] that are the coarser sediments in Quillacas. The mostly supersaturated in groundwater differences between the two drilling sites samples of the shallow aquifer within the and the variation in As concentrations in Poopó Lake basin (Paper III); however, high water wells confirm that there is a large pH values (>8), typically found in the spatial variation in the sediment properties. groundwaters within the study area, favors desorption of As from Fe oxyhydroxides The highest concentrations of As found in surfaces (Smedley and Kinniburgh, 2002). sediment cores from Condo K and Quillacas The concentration of As and Fe in the

Figure 20. Correlation between oxalate extractable As and oxalate extractable Fe leached from the oxalate step within the sequential extraction procedure; a) Condo K, b) Quillacas.

31 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03 groundwater samples does not correlate well samples B exceeded the WHO guidelines; (r<0.15), besides oxidizing conditions of most likely all water wells also have high groundwater, indicate that reductive dis- concentrations of B in the study area. solution is not the principal mechanism that There are few previous studies in Li in controls the mobilization of As in the drinking water, the most relevant, due to the shallow aquifer of the Poopó Lake basin. similarity in the environmental conditions, is The magnitude of the As extraction correla- the study carried out in the Andes of tes well with the sediment particle size. In northern Argentina (Concha et al., 2010) both investigated profiles finer grained layers where Li concentrations in drinking water release more As. This reflects the higher reached up to 1,005 µg/L while As and B weathering rate of finer sediments, facilitated concentrations reach up to 214 µg/L and by their greater reactive surface area 5,934 µg/L, respectively. The side effects of (Gustafsson et al., 2006). chronic Li intoxication causes gastro- The high amounts of Si in the groundwater intestinal pain, diarrhea edema and hypo- suggest that competition for adsorption sites thyroidism (Aral et al., 2008); however, little with Si is another important factor causing is known about the effects of long-term high As concentration in the area. The exposure to elevated Li levels through concentration of Si at 10 mg/L, combined drinking water, hence, guidelines values need with high pH values, have been shown to to be established. effectively decrease the adsorption of As High salinity and high concentrations of (Davis and Knocke, 2001; Stollenwerk et al., combined geogenic contaminants, i.e., As, B 2002). and Li in drinking water wells and their impact on the population is an issue of 6.5. Drinking water quality major concern due to the inexistence of The results of the initial investigation carried other fresh water sources. Although many out in the entire Poopó Lake basin (Paper I) people from different villages know about show that 85% drinking water wells have the toxicity of As and the risk to the health, concentrations of As above the WHO they are more concerned about the salinity guideline and the NB512 (10 µg/L) and lack of fresh water than the quality of it. (Table 1). Besides, B and Li were also present at high concentrations. Boron con- 7. CONCLUSIONS centrations exceeded the Bolivian regulation The initial investigation of this thesis project (300 µg/L) in 100% of drinking water wells was to perform an environmental assess- while twenty six of twenty seven drinking ment especially on the effects of natural water wells exceeded the WHO guidelines processes and mining activities and its (500 µg/L). Lithium concentration in drin- impact on the water resources within the king water wells were also present in high Poopó Lake basin. As a result, only one concentrations and varied in a wide range water-well was found affected by mining. It from 41.9 to 4,256 µg/L (average: is more likely that groundwater, including 584 µg/L). the area where mining activities are carried In the second stage, forty two drinking water out, is not affected by mining. However, is wells were assessed in the southern part of not possible to say the same about surface the lake basin (Paper III & IV); results show water, which is severely impacted by mining that 90% of the studied wells have As con- especially in rivers located in the north- centrations that by far exceed the WHO eastern part of the Poopó Lake basin. guideline and the NB512 regulation. Lithium On the other hand, most groundwater from in drinking water wells was also present in the shallow aquifer around the Poopó Lake high concentrations between 46 to is affected originating from natural 3,594 µg/L (average: 850 µg/L). Boron was geochemical processes, which lead to a determined only in fifteen of the forty two natural contamination with high concentra- drinking water wells; however, in all fifteen tions of As, B and Li and high salinity; water

32 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

+ - of 85% of the drinking water wells around rate Na and HCO3 at solution with conse- the Poopó Lake exceed the WHO guideline quent pH and alkalinity increase which are for As in drinking water (10 µg/L) and it has favorable for high mobility of oxyanionic As evidenced that area which is more impacted species; within the sediments, As association by natural mobilization processes is located to ferric oxyhydroxides phases which to the southern part of the Poopó Lake adsorbs/desorbs As and is potential source basin. At the same time, geothermal waters for groundwater contamination. The extre- were investigated as potential sources of As; mely heterogeneous values of EC and As main results lead to the conclusion that concentration in groundwater, could be geothermal springs in the Eastern Cordillera explained by the local variation in sedimen- are not significant sources for As contamina- tary and hydrological characteristics. tion of groundwater and surface water due to the relatively low concentrations of As, 8. FUTURE WORK which are far below of those in groundwater The present research thesis is the initial from shallow aquifers. Geothermal waters investigation about geogenic As in surface are more likely to be mixed by shallow and groundwater; moreover is the first groundwater and locally with meteoric investigation report about the occurrence of waters. combined geogenic contaminants i.e., As, B, In a second stage, the area south of Poopó Li in drinking water wells from shallow Lake was investigated, especially because of aquifers in the area of the Poopó Lake basin the high concentrations of geogenic As and in the BA. occurrence in drinking water wells. Here is For these reasons, it is important to continue concluded that groundwater from shallow the monitoring of the sampled sites and the aquifer is naturally contaminated by TEs installed piezometers and increase the num- leading to high concentrations of As, B and ber of drinking water wells for better Li. Drinking water wells are also affected by understand the extent and the spatio- high salinity, more than 90% of the drinking temporal distribution of As and other TEs. water wells by far exceeds the WHO For a better understanding of the hydrology guidelines for As. In this area, south of of the study area, it is necessary to study the Poopó Lake, B and Li occur at higher effects of groundwater discharge and to concentrations than in other parts of the determine the groundwater flow directions basin; these elements are related to volcanic and velocity and residence times. Isotopic origin typical of the study area. The main studies are necessary to help to understand sources of As are the geological formations the origin of water, water mass balance and in the surroundings of the Poopó Lake. hydrological characteristics for groundwater Coral limestone, calcareous sandstone, ande- recharge/discharge. citic to dacitic lava, evaporate and rhyolitic The study area is located in a low income lava are main rock types which play county, for this reason it is vital to develop important roles as sources of As. The and apply low-cost technologies for the re- highest concentration of As measured in mediation of As contamination in drinking rock was found in coral limestone with water and at the same time, is also important 27 mg As/kg. Physical weathering processes the search for low-As aquifers in the region. lead to rock breakdown, which later is transported as sediments by surface runoff People in this area are living exposed to As and wind to the plains. There, sediments are and other TEs for a long period of time; for major sources for groundwater As contami- this reason it is also very important perform nation through multiple geochemical proce- new investigations in order to determine the sses which lead to the release of As to extent of exposure of the population and the shallow aquifers. These geochemical mecha- health effects caused by drinking water nisms include chemical weathering of carbo- contaminated with high concentrations of nates and silicates minerals which incorpo- As and other TEs.

33 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

9. REFERENCES Aposhian, H.V., Zakharyan, R.A., Avram, M.D., Sampayo-Reyes, A., Wollen, M.L. (2004) A review of the enzymology of As metabolism and a new potential role of hydrogen peroxide in detoxication of the trivalent As species. Toxicology and Applied Pharmacology.198 327–335. Appelo, C.A.J., Postma, D. (2005) Geochemistry, Groundwater and Pollution. 2nd Ed. AA Balkema Publishers, Leiden, 655 pp. Aral, H., Vecchio-Sadus, A. (2008) Toxicity of lithium to humans and the environment-a literature review. Ecotoxicology and Environmental Safety. 70(3) 349–356. Ball, J.W., Nordstrom, D.K., Jenne, E.A., Vivit, D.V. (1998) Chemical Analysis of Hot Springs, Pools, Geysers, and Surface Waters from Yellowstone National Park, Wyoming and its Vicinity, USGS Open-File Report, pp. 98–182. Banks, D., Markland, H., Smith, P.V., Mendez, C., Rodriguez, J., Huerta, A., Sæther, O.M. (2004). Distribution, salinity and pH dependence of elements in surface waters of the catchment areas of the Salars of Coipasa and Uyuni, Bolivian Altiplano. Journal of Geochemical Exploration. 84 141–166. Baur, W.H., Onishi, B.H. (1969) Arsenic. In: Wedepohl, K.H. (Ed.), Handbook of Geochemistry. Springer Verlag, Berlin, pp. A1–A33. Bhattacharya, P., Chatterjee, D., Jacks, G. (1997) Occurrence of arsenic contaminated groundwater in alluvial aquifers from Delta Plains, Eastern India: options for safe drinking water supply. International Journal of Water Resources Management. 13(1) 79–92. Bhattacharya, P., Claesson, M., Bundschuh, J., Sracek, O. Fagerberg, J., Jacks, G., Martin, R.A., Storniolo, A.R., Thir, J.M. (2006) Distribution and mobility of arsenic in the Río Dulce alluvial aquifers in Santiago del Estero province, Argentina. Science of the Total Environment. 358 97–120. Bhattacharya, P., Frisbie, S.H., Smith, E., Naidu, R., Jacks, G., Sarkar, B. (2002b) Arsenic in the environment: a global perspective. In: Sarkar, B. (Ed.), Handbook of heavy Metals in the Environment. Marcell Dekker Inc., New York, pp. 147–215. Bhattacharya, P., Hossain, M., Rahman, S.N., Robinson, C., Nath, B., Rahman, M., Islam, M.M., von Brömssen, M., Ahmed, K.M., Jacks, G., Chowdhury, D., Rahman, M., Jakariya, M., Persson, L.Å., Vahter, M. (2011) Temporal and seasonal variability of arsenic in drinking water wells in Matlab, Southeastern Bangladesh: A preliminary evaluation on the basis of a 4 year study. Journal of Environmental Science and Health, Part A. 46(11) 1177–1184. Bhattacharya, P., Jacks, G., Ahmed, K.M., Khan, A.A., Routh, J. (2002a) Arsenic in groundwater of the Bengal Delta Plain Aquifers in Bangladesh. Bulletin of Environmental Contamination and Toxicology. 69(4) 538–545. Bhattacharya, P., Welch, A.H., Stollenwerk, K.G., McLaughlin, M.J., Bundschuh, J., Panaullah, G. (2007) Arsenic in the environment: biology and chemistry. Science of the Total Environment. 379 109–120. Birkle, P., Bundschuh, J. (2007) Hydrogeochemical and isotope composition of geothermal fluids. In: Bundschuh, J., Alvarado, G. (Eds.), Central America–Geology, Resources and Hazards. Taylor & Francis/Balkema Publishers, pp. 777–838. Birkle, P., Bundschuh, J., Sracek, O. (2010) Mechanism of arsenic enrichment in geothermal and petroleum reservoirs fluids in Mexico. Water Research. 44 (19) 5605-5617. Birkle, P., Merkel, B. (2000) Environmental impact by spill of geothermal fluids at the geothermal field of Los Azufres, Michoacan, Mexico. Water, Air & Soil Pollution. 124(3-4), 371-410. Bothe, J.V., Brown, P.W. (1999) The stabilities of calcium arsenates. Journal of Hazardous Materials. 69(2) 197–207.

34 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Bundschuh, J., Armienta, M.A., Birkle, P., Bhattacharya, P., Matschullat, J., Mukherjee, A.B. (Eds.), (2009) Natural arsenic in groundwaters of Latin America. In: Bundschuh, J., Bhattacharya, P. (series eds.) Arsenic in the Environment, Volume 1. CRC Press/Balkema Publisher, Leiden, The Netherlands. Bundschuh, J., Bhattacharya, P. (2012) About the book series. In: Ng, J.C., Noller, B.N., Naidu, R., Bundschuh, J., Bhattacharya, P. (Eds.), Understanding the Geological and Medical Interface of Arsenic, As2012. Interdisciplinary Book Series: “Arsenic in the Environment– Proceedings”. Series Editors: Jochen Bundschuh and Prosun Bhattacharya, CRC Press/Taylor & Francis, (ISBN 978-0-415-63763-3), pp. v–vi. Bundschuh, J., Litter, M.I., Parvez, F., Román-Ross, G., Nicolli, H.B., Jiin-Shuh, J., Chen-Wuing, L. López, D., Armienta, M.A. Guilherme, L.R.G., Gomez Cuevas, A., Cornejo, L., Cumbal, L., Toujaguez, R. (2012) One century of arsenic exposure in Latin America: A review of history and occurrence from 14 countries. Science of the Total Environment. 429 2–35. Bundschuh, J., Maity, J.P. (2015): Geothermal arsenic: Occurrence, mobility and environmental implications. Renewable and Sustainable Energy Reviews. 42 1214–1222. Bundschuh, J., Pérez Carrera, A., Litter, M. (2008) Distribución del arsénico en las regiones ibérica e iberoamericana (Arsenic distribution in Ibero & Iberoamerican regions; in Spanish). Editorial CYTED, Buenos Aires, 230pp. Calizaya, A. (2009) Water Resources Management Efforts for Best Water Allocation in the Lake Poopó Basin, Bolivia. PhD Thesis, Lund University, 36pp. Centeno, J.A. (2000) Arsenic-Induced Lesions. Armed Forces Institue of Pathology, Washington DC, 48 pp. Chen, C.L., Chiou, H.Y., Hsu, L.I., Hsueh, Y.M., Wu, M.M., Wang, Y.H., Chen, C.J. (2010) Arsenic in drinking water and risk of urinary tract cancer: a follow-up study from northeastern Taiwan. Cancer epidemiology, biomarkers and prevention. 19(1) 101–110. Claesson, M., Fagerberg, J. (2003) Arsenic in groundwater of Santiago del Estero, Argentina, MSc. Thesis, KTH, 59pp. Concha, G., Broberg, K., Grandér, M., Cardozo, A., Palm, B., Vahter, M. (2010) High-level exposure to lithium, boron, cesium and arsenic via drinking water in the Andes of northern Argentina. Environmental Science and Technology. 44(17) 6875-6880. Cumbal, L., Bundschuh, J., Aguirre, V., Murgueitio, E., Tipán, I., Chavez, C. (2009) The origin of arsenic in waters and sediments from Papallacta Lake in Ecuador. In: Bundschuh, J., Armienta, M.A., Birkle, P., Bhattacharya, P., Matschullat, J., Mukherjee, A.B. (Eds.), Natural Arsenic in Groundwaters of Latin America–Occurrence, Health Impact and Remediation. Taylor & Francis Group pp. 81–90. Davis, C.C., Knocke, W.R., Edwards, M. (2001) Implications of aqueous silica sorption to iron hydroxide: Mobilization of iron colloids and interference with sorption of arsenate and humic substances. Environmental Science and Technology. 35(15) 3158–3162. Du Bray, E.A., Ludington, S., Brooks, W.E., Gamble, B.M., Ratté, J.C., Richter, D.H., Soria- Escalante, E. (1995) Compositional Characteristics of Middle to Upper Tertiary Volcanic Rocks of the Bolivian Altiplano. U.S. Geological Survey Bulletin 2119, Washington, 30 pp. Dzombak, D.A., Morel, F.M.M. (1990) Surface Complexation Modelling-Hydrous Ferric Oxide. John Wiley & Sons Ltda, New York, 393 pp. García Moreno M.E. (2006) Transport of Arsenic and Heavy Metals to Lake Poopó–Bolivia, Natural Leakage and Anthropogenic Effects. PhD Thesis, Lund University, 28pp. Garcia, M.E., Persson, M., Bengtsson, L., Berndtsson, R. (2005) History of mining in the Lake Poopó region and environmental consequences. Vatten. 61 243-248.

35 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Garreaud, R.D., Aceituno, P. (2000) Interannual Rainfall Variability over the South American Altiplano. Department of Geophysics, Universidad de Chile, Santiago. Guo, T., De Laune, R.D., Patrick Jr., W.H. (1997) The influence of sediment redox chemistry on chemically active forms of arsenic, cadmium, chromium, and zinc in estuarine sediment. Environment International 23(3) 305–316. Gustafsson, J.P., Jacks, G., Simonsson, M., Nilsson, I. (2006) Mark-och Vattenkemi Teori, Department of Land and Water Resources Engineering, KTH, Stockholm. Henke, K.R. (2009) Arsenic: Environmental Chemistry, Health Threats and Waste Treatment. John Wiley & Sons Ltd, London, 588pp. Hermansson, E., Karlsson, L. (2004) Occurrence and Distribution of Heavy Metals in the Groundwater of Poopó Basin, Bolivian Altiplano. MSc. Thesis, Lund University, 92pp. Huaranca Olivera, W.R., Neumann Redlin, C. (2000) Thematic Maps of Mineral Resources of Bolivia. Uncia Quadrangle. Hydrogeological map, sheet SE-19-12 UNCIA with Explanatory note, Sergeotecmin. INE (2012) Atlas Estadístico de Municipios (Statistical Atlas of Municipalities; in Spanish) Instituto Nacional de Estadística. La Paz, Bolivia. Kapaj, S., Peterson, H., Liber, K., Bhattacharya, P. (2006) Human health effects from chronic arsenic poisoning–A review. Journal of Environmental Science and Health, Part A. 41(10) 2399– 2428. López, D., Bundschuh, P., Birkle, P., Armienta, M.A., Cumbal, L., Sracek, O., Cornejo, L., Ormachea, M. (2012) Arsenic in volcanic geothermal fluids of Latin America. Science of the Total Environment. 429 57–75. Meng, X., Korfiatis, G.P., Chiristodoulatos, C., Bang, S., (2001) Treatment of arsenic in Bangladesh well water using a household coprecipitation and filtration system. Water Research. 35 2805–2810. Mitsunobu, S., Sakai, Y., Takahashi, Y. (2008) Characterization of Fe(III) (hydr)oxides in arsenic contaminated soil under various redox conditions by XAFS and Mössbauer spectroscopies. Applied Geochemistry. 23(11) 3236–3243. Montes de Oca I. (1997) Geografía y Recursos Naturales de Bolivia (Geography and Natural Resources of Bolivia; in Spanish). EDOBOL, La Paz, 614pp. Morteani, G., Möller, P., Dulski, P., Preinfalk, C. (2014) Major, trace element and stable isotope composition of water and muds precipitated from the hot springs of Bolivia: are the waters of the spring's potential ore forming fluids? Chemie der Erde/Geochemistry. 74(1) 49-62. Mukherjee, A.B., Bhattacharya, P. (2001) Arsenic in groundwater in the Bengal Delta plain: slow poisoning in Bangladesh. Environmental Reviews. 9(3) 189–220. NB 512 (2005) Reglamento Nacional para el Control de la Calidad del Agua para Consumo Humano (National Regulation for Control the Quality of Water for Human Consumption 2005; in Spanish). Ministerio de Servicios y Obras Públicas, La Paz. 58pp. Nicolli, H.B., Suriano, J.M., Peral, M.A.G., Ferpozzi, L.H., Baleani, O.A. (1989) Groundwater contamination with As and other trace-elements in an area of the Pampa, Province of Córdoba, Argentina. Environmental Geology and Water Sciences. 14(1) 3–16. Nimick, D.A., Moore, J.N., Dalby, C.E., Savka, M.W. (1998) The fate of geothermal arsenic in the Madison and Missouri rivers, Montana and Wyoming. Water Resources Research. 34(11) 3051-3067. Panagiotaras, D., Panagopoulos, G., Papoulis, D., Avramidis, P. (2012) Arsenic geochemistry in groundwater system, In: Panagiotaras, D. (Ed.), Geochemistry–Earth's System Processes, InTech., pp. 27–38.

36 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

Parkhurst, D.L., Appelo, C.A.J. (1999). User’s Guide to PHREEQC (Version 2)–A Computer Program for Speciation, Batch–Reaction, One–Dimensional Transport, and Inverse Geochemical Calculations. US Geological Survey. Water Resources Investigations Report 99– 4259, Denver, 312pp. Pillco, R., Bengtsson, L. (2006) Long–term and extreme water level variations of the shallow Lake Poopó, Bolivia. Hydrological Sciences Journal. 51(1) 9 –114. Proyecto Piloto Oruro (PPO) (1992) Metal Contents in Lake Sediments, Totora and Myriophyllum (PPO-013). Ministerio de Desarrollo Sostenible y Medio Ambiente, Secretaria de Medio Ambiente, Swedish Geological AB, La Paz, 128pp. Proyecto Piloto Oruro (PPO) (1996a) Impacto de la Contaminación Minera e Industrial Sobre Aguas Subterráneas (Impact of Mining and Industrial Pullution On Groundwaters; in Spanish). R-BOE-9.45-9702-PPO 9616. Ministerio de Desarrollo Sostenible y Medio Ambiente, Secretaria de Medio Ambiente, Swedish Geological AB, La Paz. Proyecto Piloto Oruro (PPO) (1996b) Impacto de la Minera y el Procesamiento de Minerales en Cursos de Aguas y Lagos (Impact of Mining and Mineral Processes On Water Courses and Lakes; in Spanish). R-BOE-9.45-9703-PPO 9612. Ministerio de Desarrollo Sostenible y Medio Ambiente, Secretaria de Medio Ambiente, Swedish Geological AB, La Paz. Proyecto Piloto Oruro (PPO) (1996c) Aspectos Ambientales de los Metales y Metaloides en el Sistema Hidrológico del Desaguadero (Environmetal Aspects of Metals and Metaloids in Desaguadero Hydrologic System; in Spanish) (PPO-010). Ministerio de Desarrollo Sostenible y Medio Ambiente, Secretaria de Medio Ambiente, Swedish Geological AB, La Paz, 97pp. Quino, I. (2006) Determinación de la Calidad Fisicoquímica de las Aguas Subterráneas en la Región Norte y Este del Lago Poopó (Groundwater Physicochemical Quality Determination in the North and East Región of Poopó Lake; in Spanish). Tesis de Licenciatura, Universidad Mayor de San Andrés. Quintanilla, J. (1994) Evaluación hidroquímica de la cuenca de los lagos Uru Uru y Poopó. (Hydrochemist assessment on Poopó and Uru Uro lakes basin, in Spanish). IIQ-UMSA, La Paz, Bolivia. Quintanilla, J., Coudrain-Ribsten, A., Martinez, J., Camacho, V. (1995) Hidroquimica de las aguas del Altiplano de Bolivia. In: Ribstein, P., Francou, B., Coudrain-Ribsten, A., Mourguiart, P. (Eds.), “Eaux, Glaciers et Changements Climatiques dans les Andes Tropicales”. Bulletin de l' Institut Français d' Études Andines. 24(3) 461–471. Quintanilla, J., Ramos, O., Ormachea, M., García, M.E., Medina, H., Thunvik, R. Arsenic contamination speciation and environmental consequences in the Bolivian plateau. (2009) In: Bundschuh, J., Armienta, M.A., Birkle, P., Bhattacharya, P., Matschullat, J., Mukherjee, A.B. (Eds.), Natural Arsenic in Groundwater of Latin America, Taylor & Francis Group, London, pp. 91–98. Ravenscroft, P., Brammer, H., Richards, K. (2009) Arsenic Pollution A Global Synthesis, John Wiley & Sons Ltd., Sussex. Raychowdhury, N., Mukherjee, A., Bhattacharya, P., Johannesson, K., Bundschuh, J., Sifuentes, G.B., Nordberg, E., Martin, R.A., del Rosario S.A. (2014) Provenance and fate of arsenic and other solutes in the Chaco-Pampean Plain of the Andean foreland, Argentina: From perspectives of hydrogeochemical modeling and regional tectonic setting. Journal of Hydrology. 518: 300-316. Risacher, F., Fritz, B. (2009) Origin of salts and brine evolution of bolivian and chilean salars. Aquatic Geochemistry. 15 123–157.

37 Mauricio Rodolfo Ormachea Muñoz TRITA LWR PhD: 2015:03

Romero, L., Alonso, H., Campano, P., Fanfani, L., Cidu, R., Dadea, C., Keegan, T., Thornton, I., Farago, M. (2003) Arsenic enrichment in waters and sediments of the Rio Loa (Second region, Chile). Applied Geochemistry. 18(9) 1399–1416. Rötting, T., Mercado, M., García, M.E., Quintanilla, J. (2013) Environmental distribution and health impacts of As and Pb in crops and soils near Vinto smelter, Oruro, Bolivia. International Journal of Environmental Science and Technology. 11(4) 935 – 948. SERGEOMIN (1999) Inventariación de Recursos Naturales Renovables (Hídricos) y No Renovables (Minerales e Hidrocarburos) del Departamento de Oruro (Inventorying of Renewable and Non-Renewable Resources of Oruro Department; in Spanish). Boletín del Servicio Geológico y Minero de Bolivia. 24. Servant-Vildary, S., Mello e Sousa, S.H. (1993) Palaeohydrology of the Quaternary saline Lake Ballivian (southern Bolivian Altiplano) based on diatom studies. International Journal of Salt Lake Research. 2 69–85. Smedley, P.L., Kinniburgh, D.G. (2002) A review of the source, behaviour and distribution of arsenic in natural waters. Applied Geochemistry. 17 517–568. Sracek, O., Bhattacharya, P., Jacks, G., Gustafsson, J.P., von Brömssen, M. (2004) Behavior of arsenic and geochemical modeling of arsenic enrichment in aqueous environment. Applied Geochemistry 19(2) 169–180. Stollenwerk, K.G. (2002) Geochemical processes controlling transport of arsenic in groundwater: a review of adsorption. In: Welch, A.H., Stollenwerk, K.G. (Eds.), Arsenic in Ground Water: Geochemistry and Occurrence. Seacaucus, New Jersey, pp. 67–100. Tanaka, T. (1988) Distribution of arsenic in the natural environment with an emphasis on rocks and solis. Applied Organometallic Chemistry. 2 283–295. Tapia, J., Audry, S., Townley, B., Duprey, J.L. (2012) Geochemical background, baseline and origin of contaminants from sediments in the mining-impacted Altiplano and Eastern Cordillera of Oruro, Bolivia. Geochemistry: Exploration, Environment, Analysis. 12 3–20. Taylor, S.R., McLennan, S.M. (1985) The Continental Crust: its Composition and Evolution. Geoscience Texts. Blackwell Scientific Publications, Oxford. 312 pp. Van Damme, P.A. (2002) Disponibilidad, Uso y Conservación de los Recursos Hídricos en el Altiplano Boliviano (Availability, Uses and Conservation of Hydric Resources in Bolivia Altiplano, in Spanish). Agualtiplano, COMARH, CGIAB, Centro de Limnología y Recursos Acuáticos U.M.S.S. 70pp. Van Den Bergh, K., Du Laing , G., Montoya, J.C., De Deckere, E., Tack, F.M.G. (2010) Arsenic in drinking water wells on the Bolivian high plain: Field monitoring and effect of salinity on removal efficiency of iron-oxides containing filters. Journal of Environmental Science and Health, Part A. 45(13) 1741–1749. Welch, A.L., Stollenwerk, K.G. (2003) Arsenic in Ground Water, Geochemistry and Occurrences. Kluwer Academic Publishers, Boston, 475 pp. Whiting, K.S. (1992) The Thermodynamics and Geochemistry of As with Application to Subsurface Waters at the Sharon Steel Superfund Site Midvale, MSc. Thesis, Colorado School of Mines. WHO (1993) Guidelines for Drinking-Water Quality. (Second edition). World Health Organization, Geneva. WHO (2003) Arsenic in Drinking-Water. Background Document for Preparation of WHO Guidelines for Drinking-Water Quality (WHO/SDE/WSH/03.04/75). World Health Organization, Geneva. WHO (2008) Guidelines for Drinking-Water Quality (Third edition). World Health Organization, Geneva.

38 Hydrogeochemistry of naturally occurring arsenic and other trace elements in the central Bolivian Altiplano

WHO (2011) Guidelines for Drinking-Water Quality. (Fourth Edition). World Health Organization, Geneva. YPFB-GEOBOL (1996) Geologic Map of Bolivia (Mapa Geológico de Bolivia; in Spanish). Yacimientos Petrolíferos Fiscales de Bolivia. Other publications: TWB. The World Bank; http://data.worldbank.org/country/bolivia. (accessed November 2014). UNICEF. United Nations International Children´s Emergency Fund; http://www.unicef.org/bolivia (accessed December 2014).