Geochemistry of Trace Elements in the Bolivian Altiplano

GEOCHEMISTRY OF TRACE ELEMENTS IN THE BOLIVIAN ALTIPLANO

Effects of natural processes and anthropogenic activities

Oswaldo Eduardo Ramos Ramos

June 2014

TRITA-LWR PHD-2014:04 ISSN 1650-8602 ISBN 978-91-7595-177-5

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Cover illustration: Top left: Lake Poopó landscape; Top right: Machacamarquita mine; Botton left: sediment of thermal water; Botton center: waste material in the Tiwanaku tailing (Photograph: Prosun Bhattacharya ©, 2009); Botton right: model concept of trace elements in Poopó Basin (Figure: Oswaldo Eduardo Ramos Ramos©, 2014).

© Oswaldo Eduardo Ramos Ramos 2014 PhD Thesis KTH-International Groundwater Arsenic Research Group Division of Land and Water Resources Engineering Department of Sustainable Development, Environmental Science and Engineering (SEED) KTH Royal Institute of Technology SE-100 44 STOCKHOLM, Sweden Reference to this publication should be written as: Ramos Ramos, O. E. (2014) “Geochemistry of trace elements in the Bolivian Altiplano - Effects of natural processes and anthropogenic activities”. PhD Thesis, TRITA LWR PHD-2014:04

Geochemistry of trace elements in the Bolivian Altiplano

Dedicated to Zaret Astrid, Leonardo Fabian and Brenda Alison

iii

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Geochemistry of trace elements in the Bolivian Altiplano

FOREWORD The thesis deals with the levels of trace metals and their mobility in the Bolivian Al- tiplano and aims at an evaluation of the risk for human exposure via different pathways, such as drinking water and food items. The area is affected by long term mining activities, covering hundreds of years from precolonial time until present day. The production of metals like tin, silver, lead and zinc is still a major share of the Bo- livian economy and is likely to remain so. The thesis deals with a multitude of ma- trixes, like surface water, groundwater, agricultural soils, crops, and lake sediments in order to evaluate the risk of mobilizing metals which might threaten the human health. An extensive literature survey has been done to set the conditions in relation to notably the South American conditions. The geology of the Titicaca region, part of the Bolivian Altiplano is complex and constitutes an important background for the study. Not only the mining activities are governing the behaviour of trace metals as the mineral deposits are present in different host rocks determining the pH and redox conditions of the terrestrial and aquatic environments. Regarding trace metals in surface water sampling and analysis has been done from headwater areas, above the mining areas down to the Poopó basin. Arsenic, cadmium, manganese and zinc concentrations in surface water exceed the WHO guidelines in some of the streams. The presence of iron hydroxyoxides provides a control of the arsenic content in groundwater as they are excellent adsorbents for arsenic under oxidizing conditions which is revealed by sequential extraction. DTPA extraction has been used to assess the plant availability of trace metals in soils. The investigation of crops has revealed that arsenic and cadmium do not exceed guideline but that lead does so in beans and potatoes. The computer code PHREEQC has been used to study trace metal species and possible equilibria with solid phases. A particular magnesium arsenate has been detected in the mineralogical studies of soils. A number of approaches has been used to assess the environmental risk of trace metals such as RAC (Risk Assessment Code), Igeo (Geoaccumulation ratio) and EF (Metal Enrichement Factor). In conclu- sion is turn out that the behaviour of arsenic requires further studies in relation to the hydrogeological conditions. The bioavailability in crops of different trace metals species, inorganic and organic deserve further attention.

Prof. Em. Gunnar Jacks Stockholm, KTH, June 2014

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Geochemistry of trace elements in the Bolivian Altiplano

ACKNOWLEDGEMENTS I would like to express my sincere gratitude to Sida/SAREC for financial support, which has made this research possible and has contributed significantly to the capacity building at Universidad Mayor de San Andrés (UMSA), also to Swedish Interna- tional (SI) and International Science Programme (ISP) in Sweden and DIPGIS at Universidad Mayor de San Andrés in Bolivia for administrative assistance, coordination and management of funds and scholarship. This research was funded through the Swedish International Development Coopera- tion Agency in Bolivia (Sida Contribution: 7500707606) and also acknowledge the support of the CAMINAR project (INCO-CT-2006-032539) funded by the Interna- tional Cooperation section of the European Commission under the 6th Framework Programme. I greatly appreciate the coordination, supervision and support of Prof. Prosun Bhattacharya, Prof. Roger Thunvik and Prof. Jorge Quintanilla that contributed greatly to the success of PhD studies at KTH and UMSA, respectively. I wish to extend profound gratitude to my supervisors Dr. Ondra Sracek and Dr. Jochen Bundschuh for their scientific guidance and constructive inputs to my research papers. My gratitude also to Jyoti Prakash Maity for his revision of my thesis and discussion, without his determined support and encouragement, I could not have accomplished this thesis. I am grateful to Prof. Gunnar Jacks at KTH for vigorous work in all the stages in lab during the sequential extractions in soil, especially in the winter vacations 2009-2010, and also for providing facilities and materials used for soils studies. I wish extend my gratitude to all of my co-authors in the papers, without their scientific discussion and input, I could not reach the publications of the scientific information. My special thanks go to my friend (Mauricio, Amalia, Israel, Efrain, Mac, Elvira, and Liliana) and colleagues (Zairis, Sandhi, Dipti, Ashis, Manoj, Tipu, Annaduzzaman, Anna and Patricia) at both UMSA and KTH for the memories times in the field and lab work, meal times and relax moments that I never ever forget. Special thanks are gratefully extended to Annaduzzaman for helping me at the formatting stage of this thesis and Mauricio for the meal times and fruitful discussion on groundwater chemistry. My gratitude to Bertil, Anna, Monica for assistance in the lab analyses at KTH; and also Aira, Britt and Jerzy for great assistance in administrative as well as health mat- ters. I value the help from Prof. Carl-Magnus Mörth at the Department of Geologi- cal Sciences, Stockholm University for providing laboratory analysis. My grateful to the administration staff at Universidad Mayor de San Andrés, Faculty and Chemistry Department and Instituto de Investigaciones Químicas (IIQ) for helping me to travel abroad without any trouble, particularly to Vice-Dean PhD Ma- ria Eugenia Garcia. My gratitude to my friends and colleagues Sulema, Rigoberto, and Luis Fernando at Especialidades Químicas who through moral, spiritual support, comments and fruitful discussions during at stages of preparation of conferences, papers and this thesis. My heartfelt thanks to my parents Albertina (R.I.P.), Adela (R.I.P.), Nestor, without their guide, love and moral support I could not have been as I am today. My acknowledgement to sister (Elba), brother (Ramiro), nephews and sisters and broth- ers in law are highly appreciated. My special grateful for my mother in law Teresa Vil- larroel for take care and love to my family during the travel to Sweden. My especial thanks go to my children, Leonardo Fabian (five years old) and Zaret Astrid (one year old), for understanding, patience and moral support during the trav-

vii

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

els far away home. My sincere and profound gratitude is extended to my wife Brenda. Her patience, love and constant encouragement in all the stages of this challenge, particularly for understand me and moral support at times when I was written this work.

Oswaldo Eduardo Ramos Ramos Stockholm, June 2014

viii

Geochemistry of trace elements in the Bolivian Altiplano

LIST OF APPENDED PAPERS Paper I Ramos Ramos, O.E., Cáceres Choque, L.F., Ormachea, M., Bhattacharya, P., Quino, I., Quintanilla, J., Sracek, O., Thunvik, R., Bundschuh, J., García, M.E. 2012. Source and behavior of arsenic and trace elements in groundwater and surface water in the Poopó Lake Basin, Bolivian Altipano. Environmental Earth. Sciences 66: 793-807. (Reprinted with permission from Springer). I participated in project designing, field work, collection of the samples and analyses, data analysis and main part of writing. Paper II Ramos Ramos, O.E., Rötting T.S., French, M., Sracek, O., Quintanilla, J., Bundschuh, J., Bhattacharya, P. 2014. Hydrochemical characteristics of groundwater and surface water in the mining region of Antequera and Poopó, Eastern Cordillera, Bolivian Altiplano. Manu- script. I performed the project designing, collection of the samples in the field work, data analysis and main part of writing. Paper III Ramos Ramos, O.E., Rötting, T.S., French, M., Sracek, O., Bundschuh, J., Quintanilla, J., Bhattacharya, P. Geochemical processes controlling mobilization of arsenic and trace elements in shallow aquifers and surface waters in the Antequera and Poopó mining regions, Bolivian Altiplano. Journal of Hydrology (In review) I participated in project designing, field work, collection of the samples and analyses, data analysis and main part of writing. Paper IV Ramos Ramos, O.E., Rötting, T.S., Orsag, V., Chambi, L., Sracek, O., Quintanilla, J., Bundschuh, J., Bhattacharya, P. 2014. Total and available trace elements concentrations in soils and evaluation of uptake by crops in the mining area of the Bolivian Altiplano. Manu- script. I participated in project designing, field work, collection of the samples and analyses, data analysis and main part of writing. Paper V Cáceres Choque, L.F., Ramos Ramos, O.E., Valdez, S.N., Choque R.R., Choque R.G., Férnandez, S.G., Sracek, O., Bhattacharya, P. 2013. Fractionation of heavy metals and assessment of contamination of the sediments of Lake Titicaca. Environmental Monitoring & As- sessment 185: 9979-9994. (Reprinted with permission from Springer). I performed the laboratory work, data analysis and main part of writing.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Geochemistry of trace elements in the Bolivian Altiplano

TABLE OF CONTENTS Foreword ...... v Acknowledgements ...... vii List of Appended Papers ...... ix Table of Contents ...... xi ABSTRACT ...... 1 1 Introduction ...... 1 1.1 Rationale ...... 3 2 Research Objectives ...... 3 3 Background Literature ...... 3 3.1 Trace elements and arsenic in the environment ...... 4 3.1.1 Trace elements and arsenic in groundwater and surface water ...... 4 3.1.2 Trace elements and arsenic in soils/ sediments ...... 5 3.1.3 Trace elements and arsenic in plants...... 6 3.2 Trace elements and arsenic in Latin America and Bolivia from natural and anthropogenic sources...... 7 3.2.1 Argentina ...... 7 3.2.2 Bolivia ...... 7 3.2.3 Brazil ...... 8 3.2.4 Chile ...... 8 4 Study Area ...... 10 4.1 Study site description ...... 10 4.1.1 Poopó Basin (PB) – (Paper I)...... 10 4.1.2 Antequera and Poopó Sub-basins – (Papers II and III) ...... 10 4.1.3 Transects Coriviri, Ventaimedia, and Poopó – (Paper IV) ...... 10 4.1.4 Titicaca Basin – (Paper V) ...... 11 4.2 Paleolakes development in northern and central Bolivian Altiplano ...... 11 4.3 Geological settings ...... 11 4.3.1 Geology around the Titicaca region ...... 11 4.3.2 Oruro mineral deposits ...... 12 4.4 Climate around the Bolivian Altiplano ...... 12 4.5 Hydrology in the Bolivian Altiplano ...... 13 4.5.1 Lake Titicaca ...... 13 4.5.2 Desaguadero River ...... 14 4.5.3 Lake Poopó ...... 14 4.6 Source of salinity in shallow aquifers in Poopó Basin (PB) ...... 14 4.6.1 Poopó Basin (PB) salinity in shallow aquifers ...... 14 4.6.2 Sub-basin (Antequera and Poopó) recharge ...... 14 4.7 Formation of depth and shallow lake sediments ...... 15 4.8 Water uses in the Bolivian Altiplano ...... 15 5 Materials and methods ...... 15 5.1 Field and laboratory methods ...... 15 5.1.1 Groundwater and surface water sampling (Papers I, II & III) ...... 15 5.1.2 Soil and crop sampling (Paper IV) ...... 16 5.1.3 Lake sediment sampling (Paper V) ...... 16 5.2 Analytical experiments ...... 17 5.2.1 Determination of major anions...... 17 5.2.2 Determination of major cations and trace elements ...... 17 5.2.3 Determination of trace element in sediments ...... 17 5.2.4 Quality control ...... 17 5.3 Geochemical modeling ...... 17 5.4 Statistical analysis ...... 17 5.5 Mineralogical investigations ...... 18 5.6 Environmental index assessments for soils, crops and lake sediments ...... 18 xi

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

5.6.1 Availability ratio (AR) (Paper IV) ...... 18 5.6.2 Transfer factor (TF) (Paper IV) ...... 18 5.6.3 Risk assessment code (RAC) (Paper V) ...... 18 5.6.4 Geoaccumulation index (Paper V) ...... 18 5.6.5 Metal enrichment factor (Paper V) ...... 18 6 Results ...... 18 6.1 Groundwater and surface water hydrochemistry in Poopó Basin (PB) ...... 18 6.1.1 Groundwater hydrochemistry ...... 18 6.1.2 Surface water hydrochemistry ...... 19 6.1.3 Hydrochemical types and major ions in groundwater and surface water ...... 20 6.2 Geochemical modeling in groundwater and surface water in the Poopó Basin .... 20 6.3 Trace elements in groundwater and surface water in the Poopó Basin ...... 21 6.3.1 TEs in groundwater ...... 21 6.3.2 TEs in surface water in the Poopó Basin ...... 22 6.4 Occurrence of arsenic in groundwater in the Poopó Basin ...... 22 6.4.1 Arsenic in groundwater of Poopó Basin ...... 22 6.4.2 Arsenic in groundwater in Antequera and Poopó Sub-basin ...... 23 6.4.3 As in surface water in the Antequera and Poopó Sub-basin ...... 23 6.4.4 Distribution of major redox sensitive species (Fe and Mn) in groundwater ...... 24 6.5 Trace element concentrations in soils (Paper IV)...... 25 6.5.1 Mineralogy and soil characteristics ...... 25 6.5.2 Soil type of studied region ...... 25 6.5.3 Extraction of major trace element in soils ...... 26 6.5.4 Available trace elements extractions in soils ...... 26 6.5.5 Sequential extraction of arsenic in soil ...... 26 6.6 Trace element concentrations in crops ...... 28 6.7 Trace element concentrations in lake sediments (Paper V) ...... 28 7 Discussion ...... 28 7.1 Mechanisms controlling the hydrochemistry in the Poopó Basin (PB) (Paper I) . 29 7.2 Distinguishing hydrochemistry mechanisms in the Antequera and Poopó Sub- basins (Paper II) ...... 30 7.2.1 Discrimination between processes related to groundwater chemistry ...... 30 7.2.2 Describing the processes in surface water chemistry ...... 33 7.3 Mobilization of trace elements and As in the hydrologic system…………………..34 7.3.1 Mobilization of trace elements and As in the groundwater environment in the Poopó Basin (PB) (Paper I) ...... 34 7.3.2 Mobilization of trace elements and As in groundwater in the Antequera and Poopó Sub- basins (Paper III) ...... 35 7.3.3 Mobilization of trace elements and As in surface water (Paper III) ...... 37 7.4 Mobilization of trace elements and As in transect soils (Paper IV) ...... 37 7.4.1 Mineralogical description of soils ...... 37 7.4.2 Characteristics of trace elements in soils: trends with aqua regia extractions ...... 39 7.4.3 Characteristics of trace elements in soil: trends with DTPA - single extraction ...... 39 7.4.4 Characteristics of arsenic in soils: trends with sequential extractions ...... 39 7.4.5 Mobilization of trace elements in lake sediments ...... 39 7.5 Arsenic and other trace elements in crops ...... 41 7.6 Environmental index assessment...... 42 7.6.1 The relationship between As and TEs in water (Papers I, II &III) ...... 42 7.6.2 Index with respect to TEs in soils and crops (Paper IV) ...... 43 7.6.3 The relationship between TEs in soil and crops: assessment of transfer factor (Paper IV) .. 44 7.6.4 Sequestration and mobility of TEs in lake sediments ...... 44 8 Conclusions ...... 45 9 Future work ...... 46 10 References ...... 47

xii

Geochemistry of trace elements in the Bolivian Altiplano

ABSTRACT The occurrence of As in groundwater in Argentina was known since 1917; however, the occurrence, distribution and mobilization of As and other trace elements (TEs) in groundwater in the Bolivian Altiplano are still quite unknown. An investigation applying a geochemical approach was conducted in the Poopó Basin and Lake Titicaca to understand processes of TEs in different systems such as water, soils, crops and sediments in mining areas. In Poopó Basin, As, Cd and Mn concentrations exceed World Health Organiza- tion (WHO) guidelines and Bolivian regulations for drinking water in different places around the basin, but Cu, Ni, Pb and Zn do not. In soils, the sequential extraction methods extracted up to 12% (fractions 1 and 2), which represent < 3.1 mg/kg of the total As content, as potentially mobilized fractions, that could be transferred to crops and/or dissolved in hydrologic system. The large pool of As can be attached due to amorphous and crystalline Fe oxide surfaces (fractions 3, 4, and 5) present in the soils. Furthermore, the concentrations of As, Cd and Pb in the edible part of the crops revealed that the concentrations of As and Cd do not exceed the international

regulation (FAO, WHO, EC, Chilean) (0.50 mg/kgfw for As and 0.10 mg/kgfw for Cd), while Pb exceeds the international regulations for beans and potatoes (for

beans 0.20 mg/kgfw and for potato 0.10 mg/kgfw). In the Lake Titicaca, principal component analysis (PCA) of TEs in sediments suggests that the Co-Ni-Cd association can be attributed to natural sources such as rock mineralization, while Cu-Fe-Mn come from effluents and mining activities, whereas Pb-Zn are mainly related to mining activities. The Risk Assessment Code (RAC) indicate “moderately to high risk” for mobilization of Cd, Co, Mn, Ni, Pb and Zn, while Cu and Fe indicate “low to moderate risk” for remobilization in the water column.

Keywords: Arsenic; Bolivian Altiplano; Eastern Cordillera; Trace elements; Sur- face water and shallow groundwater; Soils and crops.

2011); b) at the time of the Spanish con- 1 INTRODUCTION quest, when Au and Ag were the precious The Bolivian Altiplano is an area with in- metals exploited in Oruro, Potosi and Co- tense human activity since the discovery of chabamba Departments from the 16th centu- mineral resources in the colonial era. At pre- ry until the late 18th century. Since the dis- sent, it has a population of 3.451.349 (~40% covery of the most famous mine Cerro Rico of the Bolivian population) included in the in Potosi during 1554, Bolivia was consid- La Paz, Oruro and Potosi Departments. ered the major Ag producer in the world Exploitation of mineral resources is the (1554-1825), until the crisis due to the eco- principal activity since the colonial era, and nomic recession and war of Bolivian inde- has a major share in the economic develop- pendence (Diaz, 2013a); and c) post colonial ment of Bolivia for more than 400 years period in the Bolivian Republic, with the (USGS & GEOBOL, 1992). Mining activi- extraction of Sb, Bi, Pb, Ag, Sn, W and Zn, ties in Bolivia have three breakthrough peri- making Bolivia as the most important global ods: a) the pre-colonial era, when Au, Ag, exporter of Sn throughout the 20th century. Cu, Hg, Pb, and Fe were extracted (Tapia, At the beginning of 1952 the Bolivian gov-

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04 ernment created the Corporación Minera de the Oruro and Potosi regions, Colquiri and Bolivia (COMIBOL) with the objective of Porco under leasing contract, and Bolivar creating massive employment for miners, under a joint venture contract with but the prices of minerals began their long COMIBOL, mainly exploiting Sn, Ag, Pb and irreversible decline, and in 1985-1987 and Zn. According to their production sta- the prices fell to their lowest point and pro- tistic, Zn, Pb, and Sn is concentrated to duction was reduced to almost zero (Tapia, 205.000 Mton, 15.000 Mton and 6.000 Mton 2011; Diaz, 2013b). Throughout its long respectively. Meanwhile, the Manquiri Min- history, most of Bolivia´s mineral producing Company a subsidiary of the Couer tion has come from a particular group of ore d'Alene Mines Corporation is developing the deposits (USGS & GEOBOL, 1992). Cur- "San Bartolomé" in Potosi, with plans to rently, Bolivia is ranked as the third largest invest US$ 220 million. San Bartolomé is producer of Sb, fourth in Sn and Zn in the oriented to the production of silver bullion world (Tapia, 2011). in metallurgical processing of surface materi- Private investments have increased in the Boliv- als that are deposited on the slopes and pe- ian Altiplano based on the strategy of new riphery of the Cerro Rico de Potosí. Finally, leasing or joint ventures between interna- the San Cristobal mine dependent on Sumi- tional companies (through their subsidiaries tomo Corporation, is the biggest mine in with native names Sinchi Wayra, Manquiri Bolivia and producing concentrates of Zn- and San Cristobal) and COMIBOL. The Ag and Pb-Ag. It is an open pit operation three companies together account for more using the latest equipment and machinery, than half of the national production and using the floating cone method (El Diario, export of minerals (70%), cooperative com- 2012). panies constitute 21%, whereas COMIBOL, High price of minerals in the current inter- which brings together state enterprises ac- national market has increased the produc- counts for 9% of exports in recent years tion, royalties and related activities in the never exceeding the 10% barrier (El Diario, mining areas in Bolivia (La Patria, 2010a). 2012; MMM, 2013). The Bolivian mining The government has increased the national company Sinchi Wayra, a subsidiary of Swit- funding and prioritized new companies in zerland's Glencore, operates five mines in the mining areas (La Patria, 2010a, b). The

Figure 1. Map of mining activities management by COMIBOL in the Bolivian Altiplano. Whites circles principal cities; solids squares, projects implemented 2006-2010; blues circles, mines belong to COMIBOL; red lines, principal roads.

Geochemistry of trace elements in the Bolivian Altiplano minerals export had increased from US$ 429 2 RESEARCH OBJECTIVES million in 2002 to US$ 3398 million in 2011 (MMM, 2012) and, as explained above, the The main objective was to improve the un- mining industry is the main activity in Oruro derstanding of trace elements (TEs) behav- and Bolivia. ior in mining areas in the Bolivian Altiplano and how geochemical processes govern their However, the previous and current mining behavior in the interaction between surface activities by private, cooperatives and na- water, groundwater (shallow aquifer) and tional companies have produced several soils and content in crops and the lake sedi- detrimental effects such as social disarticula- ments. tion, social conflicts (CIDSE, 2009; Perales, 2010), and environmental contamination The specific objectives for this study were (generation of acid rock drainage (ARD) and to: acid mine drainage (AMD)). Power imbal- i) investigate the source, extent and behavior ance between the transnational companies of dissolved As and other trace elements and the settlements and the indigenous peo- (TEs) in groundwater and surface water in ple has lead to conflicts (CIDSE, 2009; Per- the Poopó Basin (Paper I), and more specif- ales, 2010) which have been caused the loss ically in the Antequera and Poopó sub- of the quality of their environment. basins in mining areas (Paper III). Further- more, factors and processes that lead to the 1.1 Rationale mobilization of As and TEs in groundwater A large proportion of the Bolivian Altiplano and surface water were investigated is impacted by various human activities, (Paper III). which to a major extent involves mining of ii) make a geochemical assessment of the metalliferous deposits in the region, es- groundwater and surface water to evaluate pecially due to the increased mineral prices the sources of major dissolved species and during the last decades. Proyecto Piloto elucidate the processes that govern the evo- Oruro (PPO, 1993 -1996) was the pioneer- lution of natural water in the Antequera and ing and a detailed study on the distribution Poopó sub-basins (Paper II). of TEs in surface waters, soils and mining wastes in the Oruro region (PPO-03, 1996; iii) evaluate the content of TEs in soils and PPO-04, 1996; PPO-13, 1996). However, crops along three transects in relation to these studies have not reported the overall different mining activities (Paper IV). impact of TE contamination in the Lake iv) evaluate the content of TEs in shallow Poopó Basin and Lake Titicaca in the differ- and deep sediments of Lake Titicaca, using ent matrices of the ecosystem such as sur- sequential extraction (Paper V). face- and groundwater, sediments, soil and their crops produced due to mining activi- 3 BACKGROUND LITERATURE ties. Thus, it was important to carry out the During the last 200 years after the industrial present investigation which focuses on the revolution, huge changes in the global budg- study of trace elements in various systems et of critical organic and inorganic substanc- and at different scales in the Bolivian Alti- es and trace elements at the earth´s surface plano. The study would contribute to the have occurred. Soils, plants, surface water primary understanding of the distribution and groundwater have been affected by trace and mobility of the TEs in the Bolivian Alti- elements (TEs); which may inhibit biotic plano. The government environmental regu- activities and reduce the diversity of popula- lators, stakeholder, mining industry and en- tions of flora and fauna in soil. The path- vironmental regulators can use this acquired ways for transfer TEs to human beings may knowledge in the mining areas to look for include drinking water, consuming contami- environmental solutions. nated food, and/or indirectly from consuming other products (i.e. milk) (Förstner, 1995) and in mining and arid zones especial-

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04 ly the inhalation of dust (Goix et al., 2011). 3.1.1 Trace elements and arsenic in Trace elements are essential micronutrients groundwater and surface water for living organisms, in humans and also in The composition of natural water is con- plants (Kabata-Pendias & Mukherjee, 2007). trolled by both geochemical and geo- However, some of the TEs are toxic: mercu- biological processes (Morel, 1983). About ry (Hg), lead (Pb), cadmium (Cd), copper 99% of the world’s fresh available water is (Cu), nickel (Ni) and cobalt (Co); the first groundwater; a basic source of domestic, three are particularly toxic to higher animals industrial, and agricultural water supplies in and the last three are more toxic to plants many places in the world. Therefore, the than animals and are called phytotoxic chemical composition of groundwater and (McBride, 1994; Kabata-Pendias & Mukher- its contamination has become of great con- jee, 2007) at elevated levels. Potentially haz- cern recently, since the anthropogenic activi- ardous TEs to human health at elevated ties exert pressure on the quality and quanti- levels include arsenic (As), antimony (Sb), ty of water resources. beryllium (Be), cadmium (Cd), chromium Arsenic concentrations in surface water and (Cr), copper (Cu), lead (Pb), mercury (Hg), groundwater can vary by more than four nickel (Ni), selenium (Se), silver (Ag), thalli- orders of magnitude depending on the um (Tl) and zinc (Zn) (McBride, 1994). Bio- amount of available As in the source rocks available fractions of these elements are var- and the local geochemical environment iable and are controlled by specific proper- (Matschullat, 2000; Smedley & Kinniburgh, ties of abiotic and biotic media as well as by 2002). The concentration of As in unpollut- physical and chemical properties of a given ed fresh water typically ranges from 1 – 10 element (Kabata-Pendias & Mukherjee, µg/L, rising to 100 – 5000 µg/L in areas of 2007). sulfide mineralization and mining (Mandal & Trace element (TE) contamination of soil, Suzuki, 2002). Arsenic and TEs can adsorb surface water and groundwater poses a ma- onto several surfaces, particularly iron (Fe), jor environmental and human health risk. aluminium (Al), manganese (Mn) oxides and The threat that TEs exert on human and hydroxides and silicon (Si). These natural animal health is aggravated by their long- sorbent surfaces have been recognized for term persistence in the environment (Shaw, more than a 100 years and the dissolution 1990). Additionally, the manner in which a and formation of solid hydroxides and ox- TE is bound to solids influences the mobili- ides are important in the aquatic chemistry ty, bioavailability and toxicity of the element of many metals ions (Morel, 1983; Dzombak to organisms (Bacon & Davidson, 2008). & Morel, 1990). The sorption of inorganic 3.1 Trace elements and arsenic in the ions on hydrous oxides is strongly dependent on solution conditions such as pH, ionic environment strength and the presence of competing ions In general terms of TEs abundance in the (Dzombak & Morel, 1990; Smedley & Kin- upper continental crust is as follow: Cr is the niburgh, 2002; Smedley et al., 2002). In con- 18th element (35 mg/kg), Zn is the 21th ele- th trast, desorption from hydrous oxides is one ment (52 mg/kg), Ni is the 24 element of the principal mechanisms controlling (18.6 mg/kg), Cu is the 27th (37.4 mg/kg), th th mobility of the anionic TEs especially when Pb is the 34 (12.5 mg/kg), and Cd is 60 pH values in groundwater increase (Dzom- element (Wedepohl, 1995). The most exten- bak & Morel, 1990; Smedley & Kinniburgh, sively studied TE is arsenic, although it is the 2002). 20th most abundant element in the continental crust (Mandal & Suzuki, 2002; Hudson- Geochemical properties such as redox po- Edwards et al., 2004) detectable in all soils tential (Eh) and pH are the most important and sediment. Additionally, As is a ubiqui- factors controlling As speciation (Smedley & tous 14th in seawater and 12th in the human Kinniburgh, 2002), where As and others body (Mandal & Suzuki, 2002). TEs that form oxyanions (e.g. As, Se, Sb, Mo, V, Cr, U, Re) are sensible to mobiliza-

Geochemistry of trace elements in the Bolivian Altiplano tion at pH values typically found in ground- main factors influencing their concentrations water (pH 6.5 – 8.5) under both oxidizing in the soil; also factor such as climate, organ- and reducing conditions (Smedley & Kin- ic and inorganic components of soils and niburgh, 2002). The behavior of As can be geochemical characteristics (redox potential summarized in terms of three different re- (Eh) and pH) are affecting the As content in dox zones: i) shallow, oxidizing zone with soils (Mandal & Suzuki, 2002). dissolved O2, in which Fe(III) oxide and In soils, As occurs mainly as inorganic spe- hydroxides are stable and As is adsorbed; ii) cies, As(V) (arsenate) and As(III) (arsenite), intermediate, moderately reducing zone the redox potential (Eh) and pH are the without O2, in which Fe(III) oxide and hy- main factors for As speciation, but others droxides reach reductive dissolution and As such as adsorption and microbiological ac- is released and iii) deep reducing zone, tivity play important roles. However, inor- 2- where SO4 is reduced with formation of ganic As compounds can be methylated by H2S. Arsenic can co-precipitate in secondary microorganisms, producing monomer-thy- sulfides (e.g. pyrite) (Sracek et al., 2004). larsonic acid, dimethylarsinic acid, trimethy- Many toxic TEs occur in solution as cations larsine oxide and others under oxidizing (e.g. Pb2+, Cu2+, Ni2+, Cd2+, Co2+, Zn2+) conditions (Mandal & Suzuki, 2002; Litter et which generally become increasingly insolu- al., 2008); some other studies suggest that ble as the pH increases (Dzombak & Morel, the plants are able to methlylate inorganic As 1990; Smedley & Kinniburgh, 2002). The (Raab et al., 2007). solubility of TEs (cations) is strongly limited Cadmium (Cd) is classified as a chalcophile by precipitation or coprecipitation at the element, which is associated geochemically near-neutral pH in groundwater (Smedley & with Zn in sulfide minerals in rocks (Fuge et Kinniburgh, 2002). al., 1993; Garret, 2004), both having very 3.1.2 Trace elements and arsenic in soils/ similar electronic structures, electro-nega- sediments tivities and ionization energies (Fuge et al., 1993). High mobility is attributable to the Soils are formed at the land surface by 2+ weathering processes derived from biologi- fact that Cd adsorbs rather weakly on organic matter, silicate clays, and oxides at pH cal, geological and hydrologic phenomena 2+ and they show an approximately vertical < 6. Above pH 7, Cd can coprecipitate stratification produced by the continual in- with CaCO3 or precipitates as CdCO3 and fluence of percolating water and living or- Cd phosphates which thereby limits its solu- ganisms (Sposito, 1989). The mobilization of bility (McBride, 1994). TEs from rocks to soils, sediments and dis- Copper (Cu) is also classified as a chal- solution in water is dominated by both abi- cophile element, associated to sulfide in very otic and biotic processes (Morel, 1983). They insoluble minerals such as Cu2S and CuS 2+ can be incorporated into plants and animals (McBride, 1994; Garret, 2004), but Cu and became parts of food chains (Sposito, occurs in soil as a colloid and in solution 1989; Garret, 2004). almost exclusively. However, reduction of 2+ + 0 Arsenic occurs as a major constituent in Cu to Cu and Cu is possible under reducing conditions, especially if halite or sulfide more than 200 different minerals, including + elemental As, arsenides, sulfites, arsenites, ions are present to stabilize Cu . In reduced arsenates and oxides (Smedley & Kinni- soils, the mobility of Cu is very low. Most of the soil colloids (Mn, Al and Fe oxides, sili- burgh, 2002; Mandal & Suzuki, 2002). The 2+ most important As-bearing minerals are cate clays and humus) adsorb Cu strongly, increasing with pH. Organic complexes bind mixed sulfides: arsenopyrite (FeAsS); realgar 2+ (AsS), niccolite (NiAs), cobaltite (CoAsS) Cu more strongly than any other divalent (Matschullat, 2000). The terrestrial abun- transition metal; the solubility of these com- dance of As is around 1.5 – 3 mg/kg, where plexes is rather low limiting their bioavaila- parent rocks and human activities are the bility. Conversely, the mobility of soluble

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04 complexes of Cu2+ may be significant under 3.1.3 Trace elements and arsenic in plants high pH conditions (McBride, 1994). Trace element concentrations in plants re- Lead (Pb) is a strongly chalcophilic element, flect, in most cases their abundance in occurring primarily as PbS in rocks and in- growth media (soil, nutrient solution, water) soluble in reduced soils (McBride, 1994; and in ambient air. Metabolic fate and role Garret, 2004). Under oxidizing conditions, of each element in plants can be character- the Pb2+ ion becomes less soluble as the ized in relation to the basic processes such soil´s pH is raised. Complexes with organic as: i) uptake (absorption), ii) transport within matter, chemisorption on oxides and silicate plants, concentration, and speciation, iii) clays, and precipitation as the carbonate, metabolic processes, iv) deficiency and tox- hydroxide, or phosphate are all favored at icity, and v) ionic competition and interac- higher pH. In alkaline soils, solubility may tion (Kabata-Pendias & Mukherjee, 2007). increase by formation of soluble Pb-organic Trace elements such as Cu, Fe, Mn, Mo, and and Pb-hydroxy complexes. The ion Pb2+ Zn play a key role in plant metabolisms and has a particularly high affinity to Mn oxides, are constituents of several enzymes. Howev- this fact perhaps explained by Mn oxidation er, different plants exhibit various tendencies of Pb2+ to Pb4+, a very insoluble ion. Lead is in uptake of TEs, depending on several fac- the least mobile TE in soils, especially under tors, both internal and external. Also, plants reducing or non-acid conditions (McBride, exhibit a variable, and sometimes specific, 1994). ability to absorb TEs from soil and by The divalent nickel (Ni2+) is the only stable above-ground parts from atmospheric depo- form of nickel (Ni) in soil environment. The sition (Kabata-Pendias & Mukherjee, 2007). ion Ni2+ is almost as electronegative as Cu2+. The toxicity of Ni to plants occurs in acid This fact and its electronic structure favor soils formed from serpentinite or other ul- the formation of complexes with organic trabasic rocks. High organic matter levels in matter that are comparable in stability to Ni-rich soils can solubilize Ni2+ as organic those of Cu2+. As the smallest of the divalent complexes, at least at higher pH. Nickel is a ions Ni2+ fits easily into octahedral sites, co- strongly phytotoxic element, being several precipitating readily into Mn and Fe oxides times more toxic than Cu (McBride, 1994). in soils. Chemisorption on oxides, noncrys- Most of Pb in soils appears to be unavailable talline alumino-silicates, and layers silicate to above ground parts of plants, but plants are favorable above pH 6, but lower pH can absorb Pb2+ and concentrate it in their favors exchangeable and soluble Ni2+ roots, carrying very little from roots to (McBride, 1994). above ground parts as long as the plants are Zinc (Zn) in soils tends to occur as a sulfide actively growing. Toxic effects of Pb on mineral (a chalcophile mineral, sphalerite plants have not often been observed, but a ZnS) in rocks. In acid aerobic soils, Zn has hazard to animal exists because of the inher- moderate mobility, and is held in exchange- ently higher toxicity of this element to ani- able forms on clays and organic matter. At mals (McBride, 1994). higher pH, however, chemisorption on ox- Toxicity of Zn to plants is most likely to ides and alumino-silicates and complexes appear in acid soils that have not been sub- with humus decrease the solubility of Zn2+. jected to prolonged acid leaching. The rather Consequently, Zn2+ mobility in neutral and high potential solubility of Zn2+ in acid soils slightly alkaline soils is from very to ex- and the fact that Zn2+ is typically a high con- tremely low, while Zn-organic complexes centration pollutant of industrial wastes and can become soluble and raise mobility. In sewage sludge, both combine to create a strongly alkaline soils, Zn-hydroxy anions significant potential for phytotoxicity from may form to increase the solubility adding wastes of soil (McBride, 1994). (McBride, 1994). Arsenic accumulation in a plant depends on the amount available in soils. However, it can be sorbed on clay and complexed by 6

Geochemistry of trace elements in the Bolivian Altiplano organic matter which thereby decreases its Company (VCM) in Oruro. Their results bioavailability. Organic acids and root exu- suggested that all crops cultivated around date can increase the amount of available As VCM largely exceed As and Pb guidelines and exert an action on sites where the As is for crop consumption. adsorbed. Arsenite shows a high toxicity to humans compared to arsenate and organic 3.2 Trace elements and arsenic in species like arsenobetaine which is less toxic Latin America and Bolivia from (Bergqvist et al., 2014). In a recent study, natural and anthropogenic Zheng et al. (2011) showed that the root sources hairs of barley play an important role in the 3.2.1 Argentina transfer of Cd from soils. In addition, sever- In the beginning of the 20th century (1917), al studies have reported that Cd transloca- Argentina was the first country to report the tion from roots to shoots is driven by tran- presence of As in hydrologic systems in the spiration (Zheng et al., 2011). Chaco Pampean plain (Bundschuh et al., The TE ingestion in food is the one of the 2012). Arsenic and TE studies have in- main pathways for accumulating them inside creased in several ecosystems and the occur- the human bodies over time. Food contami- rence of them has become a great concern; nated with Cd can cause bone fracture, kid- actually, there are several studies in terms of ney dysfunction, hypertension, and even TEs in different ecosystems: cancer (Nordberg et al., 2002; Turkdogan et Natural sources al., 2003). Mn and Cu can cause mental dis- A thorough review of As and other TEs in eases such as Alzheimer and manganism in groundwater was performed by Nicolli et al. high concentrations (Dieter et al., 2005). Ni (2012a). They summarized the sources and ingestion can cause dizziness, fatigue, head- explained the dissolution, and the desorption ache, heart problem, fatal cardiac arrest, and processes in the Chaco-Pampean plain respiratory illness (Muhammad et al., 2011). (Smedley et al., 2005; Bhattacharya et al., Excessive Zn can cause a sideroblastic ane- 2006b; Bundschuh et al., 2012; Nicolli et al., mia; while, Zn deficiency can cause anorexia, 2010; Nicolli et al., 2012a, b; Rosso et al., diarrhea, dermatitis, depression immune 2011; O`Reilly et al., 2010). Also, As specia- dysfunction, poor wound healing and a im- tion in surface water (rivers and lakes) and pared immune response. Therefore, an ade- thermal springs sites (Farnfield et al., 2012) quate amount of Zn is very important for and TEs in the food chain (Revenga et al., normal body functions (Khan et al., 2013). 2012) and As in milk (Pérez-Carrera & Fer- In a recent review, Bundschuh et al. (2012) nández-Cirelli, 2005) have been studied. discussed the presence of As in the food Anthropogenic sources chain across Latin America. Among the countries in Latin America, in Mexico (Prie- Bermudez et al. (2012) found that the fluxes to-García et al., 2005), Argentina (Perez- of As, Cu, Pb and Zn in atmospheric depo- Carrera et al., 2009) and Chile (Sancha & sition in Cordoba were higher than in other Marchetti, 2009) have significant levels of agro-ecosystems. TEs and As has been reported in food (Díaz 3.2.2 Bolivia et al., 2004; Muñoz et al., 2002). Queirolo et Natural sources al. (2000) reported high As, Cd and Pb con- In the Bolivian Altiplano, Dejoux and Iltis tents in vegetables in Andean villages of (1992) conducted a review on Lake Titicaca, northern Chile. In other areas of Chile stud- this review was about formation and geolog- ies showed that the levels of total As in veg- ical evolution (Lavenu, 1992), hydrochemis- etables were below the permissible limit of try of the lake and inflow rivers (Carmouze the Chilean legislation (Muñoz et al., 2002). et al., 1992), climatology and hydrology Recently, Rötting et al. (2013) have studied (Roche et al., 1992) and limnology evalua- the As and Pb contents in soils and crops tion; making a whole analysis of biotic and near the smelter of the Vinto Metallurgical abiotic processes that occur in Lake Titicaca. 7

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Anthropogenic sources 2007; Tapia & Audry, 2013). Recently, Goix In the beginning of the 21st century in the La et al. (2011) carried out a study of the Plata Basin (south of Bolivia), studies on transport of atmospheric aerosols and found TEs were carried out in mining-affected high TE concentrations, some of them high- areas. Smolders et al. (2004), reported that er than values reported in the literature, con- the water chemistry of the Pilcomayo River firming that the smelters are one of the was highly variable during the year and the sources and another probably the resuspen- possible effect on biota, especially regarding sion of soil dust by ore deposit transporta- metal speciation and metal toxicity was dis- tion. Rötting et al. (2013) studied the As and cussed briefly. Miller et al. (2004) examined Pb contents in soils and crops near the smel- four communities along the Pilcomayo Riv- ter of the Vinto Metallurgical Company er, and found that the most significantly (VCM) in Oruro, their results suggesting that contaminated soils occurred with Cd, Pb, any crops cultivated around VCM should and Zn, these concentrations exceeded rec- not be consumed due to the high As and Pb ommended guideline values for agricultural contents in the crop. use. However, the results suggested that 3.2.3 Brazil most vegetables do not accumulate signifi- Natural sources cant quantities of TEs. Hence, the most significant exposure pathway appears to be TE contents in fish and seafood were evalu- the ingestion of contaminated soil particles ated (Medeiros et al., 2012; Morgano et al., attached to and consumed with vegetables. 2011), as well as in rice (Batista et al., 2010). Archer et al. (2005) concluded that mining In addition, Scarpelli (2005) evaluated As in should not be held solely responsible for As rivers of the Amazon Basin, from the Andes contamination of drinking water in the Pil- Cordillera to the Atlantic ocean. A thorough comayo basin, there is also a natural weath- study of As in different ecosystems and hu- ering of bedrocks. Kossoff et al. (2011, mans in Iron Quadrangle was performed by 2012a, 2012b) found that a high degree of Deschamps and Matschullat (2007). the metal mobility was recorded in the tail- Anthropogenic sources ing dam in spill-simulated column experi- Studies on TEs were focused on As, Ono et ments, stressing the potentially serious al. (2012) found that the total As concentra- short-to-long-term effects that accompany tion in soils was high in mining areas, how- tailing dam spill into floodplain environ- ever, bioavailable As was low (< 4.2%), indi- ment. Also, they demonstrated the effect of cating a low risk from exposure of children mine tailing on the load and cycling of As, P, next to this area. Amount of As and TEs in Pb and Sb in foodplain soils and showed surface water and sediments in the Iron that their natural weathering and oxidation Quadrangle study were assessed using a se- can lead to formation of relatively insoluble quential extraction procedure to evaluate As, Pb and Sb-bearing phases that have wide mobility of As and TEs (Varejao et al., pH – Eh stability fields. 2011). In the southern part of the Bolivian Alti- 3.2.4 Chile plano, studies at Poopó Basin had focused Natural sources on assessment of TEs in surface water, soil Díaz et al. (2004) studied the contribution of and mine waste material, the Proyecto Piloto water, bread and vegetables to the dietary Oruro (PPO–03,1996; PPO–04, 1996; intake of As in a rural village in northern PPO–13, 1996) was the pioneer study in the Chile. Also, Muñoz et al. (2005) studied the Oruro mining areas; others studies were dietary intake of As by the population in the made to understand the geochemical pro- nation´s capital Santiago, it was estimated as cesses in groundwater and surface water being 77 µg/day using a total diet study. (Banks et al., 2002; Lilja & Linde, 2006; Quino, 2006) and in lake sediments (Cáceres Choque et al., 2004, 2013; Selander & Svan,

Geochemistry of trace elements in the Bolivian Altiplano

Figure 2. a) Eastern and Western Cordillera, b) Poopó Basin, and c) Antequera and Poopó sub-basins and (T1, T2, and T3) transects scale. Location of sampling places surface water, groundwater, soil, and crops. Political borders, orange colored Poopó; pink colored, Ante- quera and magenta colored, Pazña municipalities.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Anthropogenic sources 5000 m a.s.l.; above sea level); ii) western Neaman et al. (2012) found that application mountain region, occupying only a small of earthworms in Cu and As contaminated part of the basin (3800 – 4800 m a.s.l.) and soils improves the quality of soil. Molina et finally iii) Inter-Andean flat, constituting half al. (2009) studied contribution of fertilizers of basin (3700 – 3800 m a.s.l.) (Pillco & in term of TEs and found that long-term use Bengtsson, 2006). The PB includes 22 of P fertilizer may increase levels of As, Cd ephemeral inflow rivers, the biggest are the and other TEs in agricultural soils. Marquez, Sevaruyo, Desaguadero and Ante- quera Rivers (Pillco & Bengtsson, 2006). 4 STUDY AREA The Poopó Basin has been divided into five regions (R-I – R-V) to better understand the 4.1 Study site description diverse environmental aspects (Fig. 2) Three different units are discussed i.e. basin, (PAADO, 2005). sub-basin and transect: i) Titicaca and Poopó Basins are considered in the northern and 4.1.2 Antequera and Poopó Sub-basins – southern areas of the Bolivian Altiplano (Papers II and III) (BA), ii) second, Antequera and Poopó Sub- The Poopó and Antequera Sub-basins are basins, whose rivers drain their waters into located in the Poopó Province of the Oruro Lake Poopó, which is part of the Poopó Department; they are located at 18°14’36” Basin (PB), and finally iii) transects located and 18°33’54” latitude south and 67°04’14” in the Poopó municipality. In order to avoid and 66°45’50” longitude west, respectively. misunderstanding with the names in the The sub-basins (Antequera and Poopó) are following each transect will be called T1 for situated on the eastern shore of Poopó Lake, Coriviri, T2 for Ventaimedia and T3 for and the rivers with same names discharge Poopó transects (Fig. 2). their waters into the Lake Poopó, and com- prise an area of 226.3 km2 and 109 km2, re- 4.1.1 Poopó Basin (PB) – (Paper I) spectively (Fig. 2). The Poopó Basin (PB) covers an area of about 25.000 km2 which is integrated by the 4.1.3 Transects Coriviri, Ventaimedia, and Lakes Poopó and Uru Uru, from a morpho- Poopó – (Paper IV) logic perspective the basin can be divided Transects T1, T2 and T3 are located on the into three regions: i) eastern mountain re- eastern side of Lake Poopó and are affected gion, occupying ~40% of total area (3800 – by mining activities in the trends of Eastern Cordillera. Soils and crops were collected in

Figure 3. Communication between Northern and Central Altiplano (Modified from For- nari et al., 2001; Tapia, 2011); b) location of the sediments sampling point in Lake Titicaca.

Geochemistry of trace elements in the Bolivian Altiplano three transects taking into account different The modern Desaguadero River (flow 390 topographical settings. The sites were select- km) drains from Lake Titicaca (elevation ed following criteria such as representativity, 3810 m a.s.l.; in relatively wet northern Alti- productivity, crop diversity, crop consump- plano) to shallow Lake Poopó (3685 m tion as described in Paper IV (Fig. 2). a.s.l.); in drier central Altiplano). Accor- dingly the interannual changes in the level of 4.1.4 Titicaca Basin – (Paper V) Lake Titicaca has a dramatic influence on Lake Titicaca is located in the Andes Cordil- discharge of Desaguadero River. The latter is lera at an elevation of 3812 m a.s.l., on the also enhanced by more local runoff, espe- border between Bolivia (45%) and Peru cially that of the Mauri River (mean flow 21 (55%). Its maximum depth is 281 m (average m3/s in the period 1977 – 2001) (Rigsby et depth 107 m). It consists of two bodies of al., 2005; Pillco & Bengtsson, 2006). water separated by the strait (i.e. Tiquina Strait); the larger body to the northern side is 4.3 Geological settings known as the Major Lake or Chucuito (area The Altiplano is located between the meta- 6450 km2), reaching in this part its greatest sedimentary Eastern Cordillera and the vol- depth near the Soto Island. Minor Lake or canic Western Cordillera of the Andes Huiñaymarca located toward the south co- (USGS & GEOBOL, 1992). The Eastern vers an area of 2112 km2, with a maximum Cordillera system runs from north to south depth of 45 m (Boulange & Aquize, 1981) and northwest to southeast, reaching alti- (Fig. 2). tudes up to 5000 m a.s.l. and is composed of 4.2 Paleolakes development in north- intensely folded and faulted marine sand- stones and shale of the Paleozoic and Meso- ern and central Bolivian Altiplano zoic ages (YPFB & GEOBOL, 1996; SER- The Altiplano is a vast endorheic basin in GEOMIN, 1999). These rocks were depos- central Andes of Peru, Bolivia and Argentina ited on the Precambrian basement and de- located between the Eastern and Western 2 formed by at least three orogenic cycles Cordilleras (200.000 km ). Since the early from the Paleozoic to the Cenozoic eras. Quaternary period, the Altiplano has always They are composed of intensely fractured been occupied by lakes, but these have not Ordovician limestones, quartzites, slates and always had the same size as the present day schists, with low angle faulting (USGS & lakes. Studies of these ancient lakes and the GEOBOL, 1992; YPFB & GEOBOL, main glacial stages in the Eastern Cordillera 1996) (Fig. 1a). have allowed the establishment of relation- ships between three lake formations and In the Bolivian Altiplano and Western Cor- three most recent stages of glacial recession, dillera sediment deposits corresponding to Lake Ballivian in the northern basin and the Miocene to Pliocene ages that host red- Escara, Minchin (30 – 25 ka BP; before pre- bed copper deposits and mainly epithermal sent) and Tauca (12 – 10 ka BP) in the Ag-Au-Pb-Zn-Cu deposits formed during southern basin (Risacher & Fritz, 1991; Lav- the Middle-Late Miocene and Early Pliocene enu, 1992; PPO-03, 1996; Tapia et al., 2003) (USGS-GEOBOL, 1992; Redwood 1993). (Fig. 3). The Bolivian Sn, Au-Sb and Pb-Zn belts are hosted in the Oruro Department and in the These lake systems in the Bolivian Altiplano Eastern Cordillera (Arce-Burgoa & Goldfab, is the result of evolution of a more ancient 2009). The Bolivian Sn belt continues for system which began from lower Pleistocene, 900 km east-west throughout Bolivia from with transition at the end of Pliocene from a northernmost Argentina to southernmost relatively warm climate to a cool damp cli- Peru (Arce-Burgoa & Goldfab, 2009; Tapia mate. The presence and the size of the lakes et al., 2012) are directly related to the recession of glaciers at the start of interglacial periods 4.3.1 Geology around the Titicaca region (Lavenu, 1992; Tapia et al., 2003; Rigsby et Lake Titicaca (Lakes Major and Minor) is al., 2005). located in the north part of the endorheic 11

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Figure 4. Rain distribution in the Poopó Basin. (Source: Pillco & Bengtsson, 2006). basin which formed during the Pliocene and Pomamba; those form systems of 1 km Early Pleistocene (around 3 – 2 Ma). The length (PPO-04, 1996). These zones have latter period was marked by periodic ad- been the most exploited, even to date, due vances and retreats of glaciers, and these to their high content of Sn and Ag com- result in high-amplitude fluctuations of la- pounds of easy benefit (GEOBOL, 1992). custrine basins (Lavenu et al., 1984; Servant The Poopó-Candelaria Zn, Ag-Sn belts, cor- & Fontes, 1978; Mourguiart et al., 1992; respond to Bolivian polymetallic vein type 2000). Lake Titicaca is characterized by re- deposits, associated with small sub-volcanic ceiving drains from four different geological felsic intrusions, and zoned hydrothermal formations: Volcanic, Devonian, Carbonif- alterations around igneous bodies with vari- erous and Cretaceous and lacustrine sedi- ous Sn deposits to the northeast, east and mentation, which are characterized by six southeast of Poopó (Monserrat, Challa different facies: detrital, carbonate detrital, Apacheta & others) (GEOBOL, 1992). carbonate, organo-detrital, carbonate organo-detrital, and organic facies (Rodrigo & 4.4 Climate around the Bolivian Wirrmann, 1992). Altiplano The Andes Cordillera, running continuously 4.3.2 Oruro mineral deposits from north of the equator to the southern Mineral resources are deposited in the East- tip of South America, represents a formida- ern Cordillera which includes the well- ble obstacle to tropospheric circulation and known Bolivian Sn belt, Au-Sb, and Pb-Zn acts as a climate barrier. Along its central belts (Tapia et al., 2013), that occur in the portion (15oS – 22oS), widening of the Andes study area of the Coriviri, Poopó, Candelaria produces distinctive meteorological condi- y Bolivar-Avicaya districts. The Poopó- tions called the climate of the Altiplano. The Candelaria polymetallic belt is situated at the interest in the climate of the Altiplano has boarder of the Uyuni-Uyuni fault. West of grown as its variability has a strong impact the fault large sedimentary deposits of grav- on availability of water resources in this el, sand and silt are to be found and east of semi-arid region and adjacent lowlands the fault the Eastern Cordillera’s Palaeozoic (Garreaud et al., 2003). block is situated with its siltstone, quartzite and slate composition (PPO-04, 1996). The Altiplano rainfall is largely restricted to the austral summer season (November to The Poopó and Avicaya–Bolivar districts March) especially along its southwestern consist mainly of Silurian sediments; they part, where more than 70% of the precipita- contain quartzite and slate in the Llallagua tion occurs from December to February Formation and lutites in the Uncia For- (Garreaud et al., 2003). The climate is classi- mation with more than 2 km thickness in fied as semiarid for the northern and middle this area. The main veins are Bolivar, Nane, parts, while the southwestern part is arid 12

Geochemistry of trace elements in the Bolivian Altiplano

(TDPS, 1993). Moisture availability in tropi- potential annual and monthly evapotranspi- cal lowlands is thus not a dominant regulat- ration are 1241 mm and 103 mm, respective- ing factor for Altiplano precipitation; while ly, for the period 1990 – 2010 (Zapata, the significant El Niño Southern Oscillation 2011). (ENSO) related cooling or warming of the tropical troposphere also leads to a direct 4.5 Hydrology in the Bolivian relationship between ENSO phases and Altiplano near-surface temperature fluctuations over The main hydrologic system of the Altiplano the Altiplano (Garreaud et al., 2003). Mois- corresponds to Lake Titicaca (Chucuito (Ma- ture from the Amazonian basin is advected jor) and Huayñamarca (Minor)), it is drained into the central Andes. Warm saturated air by the Desaguadero river which reaches condenses at higher altitude, producing rain- Lakes Uru Uru and Poopó, finally connected fall on the eastern slope of the Andes and on to Coipasa Salar (TDPS, 1993). the Altiplano (Tapia et al., 2003). 4.5.1 Lake Titicaca The mean annual average temperatures of Rivers flowing to the lake can be divided the Altiplano are between 7oC and 10oC into three groups with distinctive drainage (Roche et al., 1992). The mean potential basin lithologies. The first group is com- evaporation rate over the entire Altiplano is prised by the Suches, Huaycho and Keka estimated at more than 1500 mm/year Rivers, which drain from the Eastern Cordil- (Roche et al., 1992) and varies from 1500 to lera. Their drainage basins are composed 1800 mm/year in the north-south direction mostly of Silurian-Devonian continental (TDPS, 1993). However, pan observations sedimentary rocks, Permian marine sedi- and Penman equation correction have esti- ments and minor Tertiary lavas and intrusive mated the potential evaporation close to rocks. The second group includes the Coata, 1700 mm/year in Lake Poopó (Pillco & Ilave and Ilpa Rivers, which drain the West- Bengtsson, 2006). ern Cordillera with Plioceno-Quaternary volcanic and ignimbrite formations, in addi- Around Lake Titicaca the temperature re- tion to lavas, ignimbrites and associated vol- mains above 8oC and the presence of the canogenic sediments, lesser amounts of lake makes the climate more temperate de- Jurasic and Cretaceous marine sediments creasing the temperature range, but it does and Tertiary continental sediments. The not seem to cause an increase of the mean third group consists of rivers with drainage annual temperature by more than 2oC basins north of Lake Titicaca (Lake Major) around its margins. The mean annual relative and their drainage basins include Cretaceous humidity around Lake Titicaca varies be- continental and marine sediments (including tween 50% and 65% (Roche et al., 1992). limestones); Permian, Tertiary and Quater- The rainy season is from December to nary volcanic rocks and volcanogenic sedi- March and is centered in January, while the ments; Silurian, Devonian and Ordovician middle of the dry season (May to August) is marine sediments and Carboniferous sedi- June; two transitional periods separate these ments; and widespread Quarternary lacus- seasons, one in April and the other from trine deposits (Carmouze et al., 1992; Roche September to November. Rainfall over the et al., 1992; Grove et al., 2003). lakes is greater than 800 mm and can reach Higher Sr isotopic values in waters from 1000 mm. Maximum monthly rainfall rec- Lake Minor indicate a lack of complete ex- orded over the lake reached 300 to 450 mm change with Lake Major; additionnally, dif- in January 1984, a particularly rainy month ference of distribution in Sr isotopes be- (Roche et al., 1992). tween the northeast and southwest areas of Available rainfall data for the Poopó Basin Lake Minor also demonstrates limited ex- show that the average annual rainfall for a 42 change between and differing inputs to the year period (1960 – 2003) is 372 mm (Pillco three sub-basins of Lake Minor (Carmouze & Bengtsson, 2006) (Fig. 4). The median et al., 1992; Roche et al., 1992; Grove et al., 13

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

2003). At modern lake levels, about 10% of both, to explain the formation and mainte- the annual input of water to Lake Titicaca nance of these paleolakes (Grove et leaves the basin through its outlet (i.e. the al., 2003). Desaguadero River). The previous studies on the BA to deter- 4.5.2 Desaguadero River mine the origin of groundwater recharge at Lakes Titicaca and Poopó are connected by different aquifers were performed in the the Desaguadero River (390 km) and there central BA, Coudrain-Ribstein et al. (1995a, are several tributaries on its way to Lake b). Hydrogeological modeling and interpre- 18 Poopó. The main tributary is the Mauri Riv- tation of isotopic data of δ O for rain, river er located southwest of Lake Titicaca. The (Desaguadero) and groundwater indicate average annual discharge in downstream that the latter has similar values (-15‰) to Desaguadero River at the inlet to Lake Poo- those of summer rainstorms which recharge pó Basin is 66 m3/s in the period 1960 – the aquifer and the zone where recharge by 2002; the maximum and minimum annual the river Desaguadero was expected the 18 discharge is 279 and 16 m3/s, respectively groundwater δ O are higher (-10‰). Elec- 18 (Pillco & Bengtsson, 2006). trical conductivity (EC) and δ O data show that the evaporation is not sufficient to ex- Downstream from the Lake Titicaca outlet, plain the increasing salinity in the groundwa- flow in rivers is increased by seasonally vari- ter. Dissolution of salty Quaternary sedi- able discharge from tributaries that strongly ments is likely to contribute to the excess influence the hydrology and chemistry of the salinity. Additionally, isotopic studies of trit- Desaguadero River. Further downstream, ium (3H) and 14C analysis indicate that the the river becomes increasingly diluted by groundwater is 4000 years old in the basin freshwater tributaries, some of which drain located in the central BA (Coudrain-Ribstein Altiplano sediments and also volcanic rocks et al., 1995a, b). of the Western Cordillera (e.g. the Mauri River) (Grove et al., 2003). Almost all rivers disappear before reaching the alluvial fans and Lake Poopó, suggesting 4.5.3 Lake Poopó that some part of their water is reaching the The water is highly concentrated by evapora- aquifers during the dry period, while the tion in Lake Poopó and isotopically enriched opposite happening during the rainy period, by the Machacamarca, Poopó, Pazña, Sevar- with high runoff events which cause flood- uyo, Marquez and Mulato Rivers. Their wa- ing in the alluvial fans in many places around ter drains from Silurian-Devonian sand- Lake Poopó. This observation is in agree- stones and mudstones of the Eastern Cordil- ment with both conditions of clay sediments lera as well as Miocene ignimbrites and which represent low permeability and semi- dacitic lavas of the Los Frailes Cordillera arid conditions in the eastern basin and arid (Grove et al., 2003). conditions in southern-western of the PB. 4.6 Source of salinity in shallow aqui- Additionally, the water table is located close fers in Poopó Basin (PB) to the surface, which is linked to high evaporation rates, having produced extensive de- 4.6.1 Poopó Basin (PB) salinity in shallow posits of evaporates composed mainly of aquifers halite (Aravena, 1995). The hydrologic changes during the late Qua- 4.6.2 Sub-basin (Antequera and Poopó) ternary have evidenced the existence of recharge paleolakes in the central Bolivian Altiplano Zapata (2011) has reported that the isotopic (BA) which is marked by high shoreline ter- 18 2 races and lacustrine sediments. Some early composition (δ O and δ H) of groundwater studies found that the paleolakes were and surface water in the Poopó sub-basin, formed by inputs of glacial melt water to the has predominantly meteoric source (compared with the Local Meteoric Water Line- basin, recent work has focused on increased 2 18 precipitation, decreased evaporation, or LMWL; δ H= 8.3 δ O + 16.9). 14

Geochemistry of trace elements in the Bolivian Altiplano

Also isotopic results suggest enrichment of to the lack of information in the Poopó Ba- both isotopes δ18O and δ2H which are on sin (Calizaya, 2009). the evaporation line (δ2H= 4.7δ18O+40.8) reported previously by Lizarazu et al. (1987) 5 MATERIALS AND METHODS and Fontes et al. (1979). Hence, considering the few studies in the BA, it is possible con- 5.1 Field and laboratory methods clude that in whole basin there are two dif- 5.1.1 Groundwater and surface water ferent sources: i) the recharge of groundwa- sampling (Papers I, II & III) ter by evaporation influenced by surface Surface (n= 4) and groundwater (n= 28) water infiltration and ii) direct meteoric re- samples were collected around Poopó Basin, charge of the shallow aquifer. Furthermore, where groundwater samples were taken it is important to assess the shallow and from excavated wells (< 25 m depth) in deep aquifers in order to identify the water small towns and one drilling well (< 80 m resources for the near future in the BA. depth, Paper I). The water samples were 4.7 Formation of depth and shallow lake collected in the dry season (September 2007; sediments Paper I). Another set of surface (n= 22) and It is now well established that the level of groundwater (n= 30) samples were collected Lake Titicaca was at least 85 m below the in February 2009 in the rainy season in the present day level at some time during the Antequera and Poopó Sub-basin (Papers II early and middle Holocene and higher than & III); most of the surface water was located the modern level during the Last Glacial in the headwater region (headwaters: i.e. Maximum (LGM; Baker et al., 2001). The before flowing through areas with mining LGM sediments were influenced globally activities or the whole sub-basin). which is documented by oxidized units else- A Nansen bottle (plastic material) was used where in the world under probably similar to take groundwater samples according to a conditions (McArthur et al., 2008). Around procedure describing by standard sampling the Bolivian Altiplano the LGM (ca. 18–20 methods (Bhattacharya et al., 2006a, b). The ka) caused the absence of the ostracods, sample bottles were rinsed three times with which has been interpreted as unfavourable water at each sampling site. Samples were conditions for calcified organisms because collected and filtered through 0.45 µm filters the climate was much colder and the pH was using a hand filtration unit; they were divid- lower (Mourguiart et al., 1992; Mourguiart, ed in three lots: 2000). This global condition in the sedi- Type 1: 100 mL for major anion analysis; for ments has affected the biogeochemistry - - 2- chloride (Cl ); nitrate (NO ); sulfate (SO ). around the Altiplano. 3 4 Type 2: 100 mL for major cation analysis: 4.8 Water uses in the Bolivian Alti- sodium (Na+); potassium (K+); calcium plano (Ca2+); and magnesium (Mg2+). In the Titicaca Basin, currently, there are Type 3: 100 mL for TEs and phosphate many small companies concerned with fish 3- farming and tourism in the Lake Major (PO4 ); filtered samples acidified with which are increasing their activities, and also HNO3 (pH< 2) (aluminum (Al); arsenic the inhabitants like fishing of native and (As); cadmium (Cd); copper (Cu); lead (Pb); introduced species. iron (Fe); manganese (Mn); silicon (Si); zinc (Zn)). All samples were stored at low tem- In the Poopó Basin, the evaluation of water perature (4oC) until further analyses in lab. uses has estimated that the major consumption is for industrial (included mining) activi- Multiparameter equipment was calibrated ties 0.5 – 0.6 m3/s, for irrigation 8.1 m3/s, using standard solutions for pH, electrical and for domestic use 0.2 m3/s. The mining conductivity (EC), redox potential (Eh) and consumption could be underestimated due total dissolved solid (TDS) in the field and

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04 measurements were performed as described residual phases, the method is described in in Papers I, II and III. detail in Paper IV. Alkalinity was also determined in the field by The TE concentration associated to a specif- titration with 0.01 M HCl until the endpoint ic extractant was expressed as follows: [i]j (at pH 4), as indicated by a color change of a where i is the TE and j relates to extractant mixed acid-base indicator (indigo carmine (i.e., [i]regia for HNO3/HCl, [i]DTPA for DTPA and methyl orange; Merck proanalysis rea- and [i]stp1-stp5 for sequential extraction proce- gent) (Papers I, II & III). dure in the step 1 – 5). 5.1.2 Soil and crop sampling (Paper IV) Crop sampling (Paper IV) Soils sampling (Paper IV) Crops (alfalfa (Medicoga sativa), barley (Horde- Top soils were collected (~2 kg) up to a um vulgare), beans (Vicia faba) and potatoes depth of 0.30 m and homogenized in the (Solanum tuberosus)) were collected during the field as composite samples. Samples were period of harvesting jointly with the soils at dried at ambient temperature and sieved the sample sites. Along each transect the through a 2 mm mesh prior to analysis in same four crops were sampled in order to lab. Physical and chemical properties were make a comparison between the places. Ed- measured in field and in lab, variables in- ible part of the crops were selected and dried clude: depth, slope, soil water content, soil at ambient temperature for a day followed o pH, texture, electrical conductivity (EC), and by subsequent oven drying at 70 – 80 C for standard methods were used to determine 24 h to remove moisture (Paper IV). + + 2+ 2+ the cations (Na , K , Ca , Mg ), organic A mixture of ultra-pure H O (30% v/v), matter (OM) (Walkley, 1946; Jackson, 1958), 2 2 HNO3 (65% m/v) and HCl (37% m/v) was cation exchange capacity (CEC), available used to digest the samples. Pseudo total con- phosphorus (P), and clay content. tents of TEs (i.e. As, Cd, and Pb) were as- For the pseudo total digestion (i.e. As, Cd, sessed, the procedure is explained in detail in Cu, Pb and Zn), the soil samples were di- Paper IV. gested in a microwave closed system (3000 5.1.3 Lake sediment sampling (Paper V) Anton Paar) with aqua regia (1:3 mixture of The samples were taken from 8 – 156 m HNO and HCl, both purified by sub- 3 depth in the Lakes Minor (Huañaymarca) boiling distillation). The residues represent and Major (Chucuito) (Lake Titicaca). They the presence of SiO and were separated by 2 were collected using a stainless steel grab filtration through Whatman filter paper; the sampler and stored in closed plastic vessels, procedure is explained in detail in Paper IV. and dried at room temperature. The samples Furthermore, in order to evaluate the bioa- were ground and passed through a 270 mesh vailable contents of TEs (i.e. As, Cd, Cu, Pb sieve (<53 µm); finally, they were homoge- and Zn) in agricultural soils, the soil samples nized and stored in polyethylene plastic bags were extracted with diethylenetriami- (Paper V). nepentaacetic acid (DPTA) (mixture of The dry samples (200 mg) were transferred DPTA, CaCl , and TEA (HOCH CH ) N) in 2 2 2 3 quantitatively to a quartz vessel and digested order to recognize potential bioavailable with aqua regia (HNO3 (65% m/v) and HCl TEs (ASA, 1982; Paper IV). (37% m/v), Merck reagents, both purified For a detailed evaluation of As in the soils a by sub-boiling distillation), and microwave sequential extraction procedure (Wenzel et (Anton Paar Multiwave closed system) was al., 2001) was applied to obtain the following used to pseudo total digestion of trace ele- fractions: non-specifically sorbed, specifically ments (i.e. Cu, Cd, Co, Fe, Mn, Ni, and Zn). sorbed, amorphous and poorly-crystalline All the samples showed almost total diges- hydrous oxides of Fe and Al, well- tion, except for some presence of SiO2. The crystallized hydrous oxides of Fe and Al and procedure is explained in detail in Paper V.

Geochemistry of trace elements in the Bolivian Altiplano

For the speciation of TEs sequential extrac- for high content of TEs (heavy metals), and tion was applied; however, sequential extrac- to determine lower concentrations, electro- tion can only indicate the potential, rather thermal atomization in a graphite furnace, than actual, mobility of TEs bound to soil HGA-800, was used at IBTEN laboratories and sediment (Ure, 1991; Bacon & Da- (Paper V). vidson, 2008). This approach has been ap- 5.2.4 Quality control plied to both sediment and soil samples as Quality control was performed for all anal- described in Papers IV and V. The Commu- yses (water: surface and groundwater, soil, nity Bureau of Reference (BCR) scheme was crop and sediment samples), which includes applied to the Lake Titicaca sediment sam- both blank (blanks, spike blanks) and sample ples (Paper V) and they were digested fol- (duplicate and spiked) control (Papers I, II, lowing the BCR sequential extraction proce- III & IV). dure (Ure et al., 1993), obtaining the fractionation of exchangeable, acid soluble, re- 5.3 Geochemical modeling ducible, oxidizable, and residual fractions Aquachem software (v4.0) was used to build (slightly modified in the last step). a water chemistry database and PHREEQC 5.2 Analytical experiments (Parkhurst & Appelo, 1999) modeling with MINTEQ.v4 database (Allison et al., 1990). 5.2.1 Determination of major anions Thermodynamic data for Ca- arsenate min- For major anion analysis in groundwater and erals and complexes were added from Bothe surface water samples ion chromatography and Brown (1999) and data for Mg com- was used at the Instituto de Investigaciones plexes from Whiting (1992) were used to Químicas (IIQ), Universidad Mayor de San explain the possible formation of complexes Andrés (UMSA) (Paper II) and at the Land and minerals related to Ca and Mg arsenate: and Water Department at KTH (Paper III) i) Calculation of saturation index (SI) was laboratories (Papers I, II, & III). made to predict and identify mineral phases 5.2.2 Determination of major cations and which are in equilibrium with aqueous phas- trace elements es in both surface and groundwater (Papers I, II & III). Saturation index is defined as For major cations (Na+, K+, Ca2+ and Mg2+) follow: when SI= 0 a minerals is at equilibri- in the groundwater and surface samples um with the aqueous phase; when SI> 0, the were determined with flame atomic absorp- water is oversaturated with respect to a min- tion spectrometer (AAS, Perkin Elmer Ana- erals and finally, when SI< 0, water is under- lyst 100) at the laboratory of the IIQ at Uni- saturated with respect to a mineral versidad Mayor de San Andrés (UMSA, Bo- (Parkhurst & Appelo, 1999). ii) Calculation livia) (Papers I, II, & III). of speciation was completed to understand TEs and As concentrations in groundwater, As species distribution of mass at equilibri- surface water, digested soils and crops were um in aqueous phases among complexes and determined using inductively coupled plasma redox couples (Papers I, II & III). iii) Calcu- optical emission spectrometry (ICP-OES; lation of speciation taking into account TEs Varian Instruments, model Varian Vista Pro in equilibrium in aqueous phases which con- Ax) at the laboratory of Department of trol the mobilization of As (Papers I, II & Geological Sciences, Stockholm University III). (Papers I, II, III & IV). 5.4 Statistical analysis 5.2.3 Determination of trace element in For bivariate statistics, Pearson´s correlation sediments matrix was carried out in Papers I, III, IV The metal concentrations in the sediment and V; additionally, Spearman´s rank corre- extracts obtained at each step were deter- lation was measured in Paper IV. Multivari- mined using an atomic absorption spectro- ate statistics approach was involved to un- photometer Perkin Elmer AAnalyst-100; derstand the principal component analysis flame atomization (air/acetylene) was used (PCA) and hierarchical cluster analysis 17

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

(HCA) which were performed in Papers I (Passos et al., 2010) and the values are inter- and V. Finally, extracted factors and eigen- preted in accordance with the RAC classifi- values and factor analysis (FA) were calcu- cation (Paper I). The range values were kept lated in Paper II. The procedures are ex- in order to make a comparison with other plained in detail in the papers. studies; nine (deep and shallow) sediments were included in the assessment. 5.5 Mineralogical investigations The identification of minerals in selected 5.6.4 Geoaccumulation index (Paper V) soils and profile samples was carried out by The geoaccumulation index (Igeo) is used to powder X-Ray Diffraction (model Seifert determine the quantitative extent of TE pol- XRD 3003 TT, with Cu Kα1 (1.5418Å) radi- lution in sediments, this index is divided into ation) to identify bulk mineral composition seven classes (Müller, 1981; Paper V). The at the Instituto de Geología y Medio Ambi- geochemical background value in average ente (IGEMA) at the Geology Department shale was used to calculate the Igeo index, of Universidad Mayor de San Andrés nine sediments samples were taken in the (UMSA) in La Paz, Bolivia. evaluation of Igeo. 5.6 Environmental index assess- 5.6.5 Metal enrichment factor (Paper V) ments for soils, crops and lake The Enrichment Factor (EF) is a ratio be- sediments tween real and reference values and is an In order to evaluate in detail the TE data estimation of anthropogenic input. The EF base in different part of the ecosystems their values were calculated with respect to refer- environmental risk values were calculated, ence background concentrations relative to explanations are included in the respective Fe concentrations, and are classified on the papers. basis of the five ranges; Paper V). In this study, the EF was calculated with respect to 5.6.1 Availability ratio (AR) (Paper IV) reference background concentrations of Fe, Availability ratio (AR) is used to estimate the since the TEs are sorbed on its surface and soil metal enrichment, or even a slow trans- act as a quasi-conservative tracer of the nat- formation to more stable forms that has ural metal-bearing phases in the sediments been discussed in previous studies (Massas (Schiff & Weisberg, 1999; Turner & Mill- et al., 2009, 2010, 2013). Twenty seven sam- ward, 2000; Aprile & Bouvy, 2008). Nine ples paired between aqua regia and DPTA sediments were taken to calculate the EF extracts were evaluated (Massas et al., 2009). values. 5.6.2 Transfer factor (TF) (Paper IV) The transfer factor, TF soil-to-plant (edible 6 RESULTS part), is a measure of the uptake of trace 6.1 Groundwater and surface water elements by plants from soil (Cui et al., hydrochemistry in Poopó Basin 2004; Rötting et al., 2013). The total of the (PB) samples (n= 36) was included in the evaluation of TF, which included beans, barley, 6.1.1 Groundwater hydrochemistry potato and alfalfa; the soil and crop concen- The field data set around the Poopó Basin trations were measured on a dry weight ba- (PB) shows that the pH of the groundwater sis, but the crop contents were converted to is nearly neutral (7.2) in the samples without fresh weight basis assuming a mean water contamination (range 6.40 – 7.90); except content of 90% (Rötting et al., 2013). the sample (29), SORP1 well, located in the mining affected area (R-II) with pH 3.79 in 5.6.3 Risk assessment code (RAC) (Paper V) the dry season (Paper I). However, a detailed The RAC is defined as the fraction of TE groundwater survey of the Antequera and exchangeable and/or associated with car- Poopó sub-basins (regions II & III) indicate bonates (%F1), this portion of the TE is that pH values are noticed between 6.6 – 9.7 easily released to the hydrologic system (median, 8.1) in groundwater and 7.7 (medi-

Geochemistry of trace elements in the Bolivian Altiplano an value) for thermal springs samples (Pa- 6.1.2 Surface water hydrochemistry pers II & III) in the rainy season. The surface water samples located in the The Eh had values in the groundwater in the headwater region(reported in Papers I, II & range 142 to 218 mV along the PB indicat- III) show that a sample in the Antequera ing moderately oxidizing conditions in all sub-basin has a pH of 7.60 in dry period; regions in the dry season (Paper I). Whereas however, in the rainy season the pH range is in the rainy season in the Antequera and from 8.05 – 12.2 taking into account the Poopó sub-basins, the ranges are -86.6 to samples in both the Antequera and Poopó 203.7 mV (median, 64.5 mV) in groundwater sub-basins. All the surface stream samples and 28 to 142.7 mV (median 86.0 mV) in located in the headwaters of the basin are thermal springs. The lowest Eh values are in dominated by baseflow. In contrast, along non-thermal groundwater samples GW-15 (- the area affected by acid mine drainage 8.3 mV), GW-27 (-12.8 mV), and GW-28 (- (AMD) and acid rock drainage (ARD) the 86.6 mV), where the Eh values suggested pH range does not show any significant dif- moderately reducing conditions in some ference between dry and rainy season (3.1 – places, but oxidizing in others (Papers II & 4.7; 3.3 – 5.2), respectively, in the Antequera III). sub-basin; instead the range is 3.7 – 7.2 in the Poopó sub-basin in the rainy season The EC values indicate different behaviors (Papers II & III). in the PB, groundwater EC values varies in all regions, for R-I from 0.9 – 2.1 mS/cm, In surface water, comparison of Eh values for R-II from 0.7 – 5.6 mS/cm, for R-III for both dry and rainy seasons in the Ante- from 0.2 – 1.2 mS/cm, for R-IV from 0.6 – quera and Poopó sub-basins shows that the 1.3 mS/cm, and from 0.6 – 3.8 mS/cm at R- range does not seem to changes and the V in the dry season (Paper II). But, during range is 117 – 381 mV (Papers I, II & III) the rainy period in the Antequera and Poopó in areas affected AMD/ARD). The samples sub-basins, the range of EC values in along the river affected by AMD/ARD groundwater is 0.13 – 3.50 mS/cm, except (from Bolivar tailings to outlet of the sub- GW-5 (13.47 mS/cm) and thermal spring basin) could be governing the oxidation of samples show the range 10.26 – 19.13 the solid material in the rivers bed during mS/cm (median, 17.24 mS/cm) (Papers II both dry and rainy seasons. Nevertheless, & III). the samples located in headwaters show that most of the surface water samples have

Figure 5. Piper plot showing the major ion chemistry of the water samples in: a) groundwater (wells) and b) surface water. Arrows in the plot indicate the possible pathway of evolution of groundwater chemistry.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04 moderately reducing or reducing conditions water which shows high EC which is con- (Papers II & III). sistent with the field data. The river baseflow in headwaters shows low The surface water is predominantly of the

EC values 0.28 mS/cm (Paper I) and in the Ca–SO4 type at R-II, except for the sample range from 3.3 µS/cm to 0.51 mS/cm (Pa- CHAO1 which is of the Ca–HCO3 type per II & III) in the dry and rainy seasons, located at the headwater area (Paper I). In respectively. However, occurrence of high the rainy season, the samples from the EC in surface water in the area affected by Antequera and Poopó sub-basins show vari- ARD/ARD is directly related to the dissolu- ous hydrochemical types: most samples fall tion of minerals and, additionally, the com- into the Na–Cl, Ca–Mg–HCO3–Cl, and Ca– bination with thermal springs to reach values Mg–CO3 water type while the Ca–SO4 water from 2.4 to 15.9 mS/cm in both seasons and type (Fig. 5b; Paper II) coincides with af- sub-basins (Papers I, II & III). fected areas by AMD/ARD. 6.1.3 Hydrochemical types and major ions in 6.2 Geochemical modeling in groundwater and surface water groundwater and surface water in The Piper diagram for groundwater in the the Poopó Basin dry season around the PB, for region I (R-I; The results of PHREEQC geochemical Fig. 5a) shows that the major ions in the modeling found that the ranges of calculated groundwater are Mg–Na–HCO3; for R-II log pCO2 values in the groundwater is -2.7 – the dominant ions seem to be Ca–HCO3, -1.5 (median, -1.9) during the dry season while others are Na–Cl and in the region around the PB; while in the rainy season affected by mining activities shows Na–SO4 samples from the Antequera and Poopó water type. In the regions R-III, R-IV and R- sub-basins indicate a range from -5.2 to -1.9 V (Fig. 5a) Na–HCO3 are the dominant ions (median, -2.9). Furthermore, during the dry in the groundwater samples followed by Na– season the samples are undersaturated with Cl (Paper I). During the rainy season in the respect to calcite and dolomite, the opposite Antequera and Poopó Sub-basins, the is true for both calcite and dolomite in the groundwater does not have a clearly pre- rainy period (Papers II & III), except for few dominantly water type; however, still Na–Cl groundwater samples in both sites. predominate in both sub-basins (Papers II & III). The enrichment of Na+ and Cl- in the The distribution of the SI values calculated groundwater samples could be linked to by the PHREEQC indicate that water was halite dissolution or the influence of thermal between equilibrium and oversaturation with

Figure 6. Piper diagrams showing hydrochemical types of a) groundwater (solid diamond Antequera, grey diamond Poopó) and b) surface water (solid squared Antequera, open squared Poopó).

Geochemistry of trace elements in the Bolivian Altiplano

- 2- respect to ferrihydrite (Fe(OH)3) in the dry HCO3 and CO3 dissolution in groundwater season (Fig. 7a; Papers I & III). However, in which could be due to carbonate (calcite or the rainy season, groundwater samples are dolomite) weathering and less associated to undersaturated with respect to rhodo- organic matter oxidation (Mukherjee et chrosite (MnCO3) and siderite (FeCO3) min- al., 2008; Papers I & II). erals (Fig. 7b, c; Paper III). Moreover, SI 6.3 Trace elements in groundwater and values indicate that water is undersatudated surface water in the Poopó Basin with respect to scorodite (FeAsO4:2H2O) during both dry and rainy seasons (Paper 6.3.1 TEs in groundwater III). In groundwater, the concentrations of Cu, In surface water, further evaluation of the SI Cd, Pb and Zn are low in the regions around results in Paper III showed that the samples the Poopó Basin (PB), except in region II in of headwater from both the Antequera and the dry season (Paper I). A detailed evalua- Poopó sub-basins are oversaturated with tion in the mining areas (region II) found respect to ferrihydrite. In contrast, in the that the TE concentration (comparison of areas affected by AMD/ARD the SI results median values) in groundwater follows the suggest that the samples are undersaturated order Zn> Fe> Mn> As> Cu> Cd in the at Antequera sub-basin, while in the Poopó Antequera sub-basin, and Mn> Zn> As> sub-basin they are oversaturated with respect Fe> Cu> Cd in the Poopó sub-basin (Paper to ferrihydrite (Fig. 7c; Paper III). III). The Cd concentrations exceed the The results from both dry and rainy seasons WHO guideline (i.e. 3.0 µg/L) in the Ante- indicate that the low values are related to quera sub-basin samples, while in the Poopó

Figure 7. Bivariate plots showing the dependency of the calculated Saturation In- dices values for ferrihydrite versus Fe in groundwater, b) siderite versus Fe in groundwater, c) rhodochrosite versus Mn in groundwater, and d) ferrihydrite versus Fe in surface water. Equiline shown at zero.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

1.68 mg/L), Mn (median 12.05 mg/L), Zn (median 116.5 mg/L) and Al (median 24.4 mg/L) in the area affected by AMD/ARD. The surface waters in this affected area show low As concentrations (<15 µg/L) (Paper I). The TE concentrations are higher in the surface water in Antequera sub-basin due to the high acid pH of the river water. In a detailed study of the surface water (Fig. 9), the order of concentration of the TEs in surface water (comparison of median values) was Zn> Fe> Cd> Mn> As> Cu in the Antequera sub-basin (Paper III). This high- lights the severity of the Cd contamination, which could be explained by the huge amoun of mining wastes abandoned along the Antequera River (e.g. at Avicaya, Fig. 1). The order of concentration for the Poopó sub-basin is Fe> Zn> As> Mn> Cd> Cu (Paper III). The headwater samples (SW-2, 3, 13 – 18) show that the Cd concentrations are between b.d.l. (<0.46 µg/L) and 2.05 µg/L. In the places where the samples are influenced by thermal water the Cd concentration ranges from 0.58 to 2.22 µg/L; all of the samples represent a mix of thermal water and river baseflow. The samples influenced by ARD/AMD along the Antequera and Poo- pó rivers are typical of mining areas with high TE (Al, Cd, Cu, Mn, Pb and Zn) con- Figure 8. Box and Whisker plots for centrations during both dry and rainy sea- metals and TEs in groundwater regions; sons (Paper III); furthermore, in a sample a) region I, b) region II, c) region III, taken near to the Tiwanaku tailing dam (SW- d) region IV, e) region V and f) surface 22) indicates high Cd concentration in the water. range 7.3 – 1855 µg/L (median, 155.9 µg/L), Pb (239.3 µg/L) and huge concentrations of sub-basin samples show less Cd (median, Fe and Mn (Fig. 9c, d; Paper III). 0.95 µg/L). For Cu the range is 1.0 – 8.15 µg/L (median, 2.8 and 2.4 µg/L in the 6.4 Occurrence of arsenic in Antequera and Poopó sub-basins, respec- groundwater in the Poopó Basin tively). However, Zn concentrations are in 6.4.1 Arsenic in groundwater of Poopó the range 6.5 – 1318 µg/L (median, 107.5 Basin and 25.2 µg/L in the Antequera and Poopó In the whole Poopó Basin (PB), the As con- sub-basins, respectively) during the rainy centration in the groundwater ranges from period (Fig. 8a – e; Paper III). below detection limit (b.d.l., 5.6 µg/L) to 6.3.2 TEs in surface water in the Poopó 242.4 µg/L (median, 41.4 µg/L; n = 28) Basin taking into account the whole database. In The surface water samples (Fig. 8f; Paper I) order to better understand the As distribu- indicate high concentration of Fe (median tion the area was divided into five regions

Geochemistry of trace elements in the Bolivian Altiplano

Figure 9. Box and Whisker plot showing the distribution of As and trace elements in a) Antequera groundwater, b) Poopó groundwater, and c) Ante- quera surface water, d) Poopó surface water. based on previous studies (PAADO, 2005). 364 µg/L. The study has taken into account In the groundwater samples for R-I in the three limits. First, WHO and Bolivian regu- northern of the basin (Paper I) the results lation in terms of safe drinking water guide- for As concentration are relatively low (b.d.l. lines (WHO & NB-512) up to 10 µg/L; sec- to 24.3 µg/L; Fig. 8a). However, for R-II in ondly, the values of the maximum content in the eastern area (Paper I) with mining activi- the Bolivian Class A standard for discharge ties the concentration of As ranges from to water bodies (50 µg/L) and, third, con- b.d.l. up to 185.7 µg/L (Fig. 8b) reaching centrations of As above the later limit. In the maximum concentration in well PAZP2 Antequera sub-basin, two samples are below (185.7 µg/L). As concentrations in region III the first limit, eight samples are below the (southern of Poopó Basin) are higher than in second limit (< 50 µg/L) and seven exceed other regions, these are located near the the Bolivian limit (>50 µg/L) reaching 364 Condo K and Quillacas towns, reaching µg/L as highest value. Meanwhile, in the 229.4 and 233.3 µg/L, respectively. In R-IV Poopó sub-basin, As concentration are in (Paper I) the As concentrations range from the range from >10 µg/L to 104.4 µg/L. 38.8 to 116.8 µg/L in the eastern side of the Eleven samples do not exceed the second basin. Finally, in R-V near Toledo located in limit, and only one sample shows high con- the northwest, the highest As concentrations centration than 50 µg/L (i.e. 104 µg/L). are 234.3 – 242.4 µg/L in the dry season Moreover, the As concentration in thermal (Fig. 8a – e; Paper I). springs is in the range 34.1 – 45.2 µg/L (median value, 38.9 µg/L) in both the Ante- 6.4.2 Arsenic in groundwater in Antequera and Poopó Sub-basin quera and Poopó sub-basins (Paper III). In the detailed assessment of the rainy sea- 6.4.3 As in surface water in the Antequera son in the Antequera sub-basin, the As con- and Poopó Sub-basin centration in groundwater samples (Fig. 9a, The As concentration in surface water in the b) are in the range from b.d.l. (5.6 µg/L) to Antequera sub-basin (Fig. 9c, d) that is im- 23

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

F i g u re 10. X- ray diffraction of selected soils samples: a) three transect (T1, T2, and T3) and profile POM2 (4, 7 and 9 m; deep), and b) standard magnesium arsenate mineral (Mg 2As2O7). pacted by ARD/AMD has relatively low As Antequera and Poopó sampling sites. concentrations (median, 9.7 µg/L), presum- PHREEQC modeling results of speciation ably because of As immobilization due to suggest that Fe(II) concentrations in all sur- co-precipitation/adsorption on ferrihydrite face water samples are greater than Fe(III) in surfaces, in the dry season (Paper I). The areas affected by ARD/AMD. Conversely, headwater sample of the sub-basin shows Fe(III) concentrations predominate over 32.4 µg/L of As. Fe(II) in the samples in the headwaters of During the rainy period in the Antequera the basin. sub-basin, two samples were found to have 6.4.4 Distribution of major redox sensitive As concentrations between b.d.l. and 10 species (Fe and Mn) in groundwater µg/L, and ten samples fell in the range (10 – In the entire basin, the concentrations of Fe 50 µg/L), i.e. between the WHO/NB-512 and Mn groundwater are in the ranges 9.8 – guideline and the Bolivian regulation (50 1,892 µg/L (median 22.5 µg/L), and 1.2 – µg/L). One sample exceeds the Bolivian 1,089 µg/L (median 7.3 µg/L), respectively limit as it contains 248 µg/L As (SW-8). In (Paper I). High concentration of Fe and Mn the Poopó sub-basin, two samples have As are found at R-II in the alluvial plain and the 2- concentrations between b.d.l. to 10 µg/L, SO4 concentration is relatively low (9.1 seven samples fell in the range (10 – 50 mg/L, Paper I). The sources for increased µg/L) and one sample exceeds all regula- concentrations of these TEs may be wolf- tions (i.e. 30.7 As mg/L; SW-22, tailing leak- ramite ((Fe,Mn)WO4) veins reported by age). The samples in the headwater of the Banks et al. (2004). However, low concentra- sub-basins show concentrations in the range tions of these TEs are found in the other from b.d.l. to 34.3 µg/L As (Paper III). regions (R-I, R-III, R-IV and R-V) with few PHREEQC results suggest that As is ad- exceptions (Fig. 8a – e; Paper I). sorbed more readily onto ferrihydrite in both

Geochemistry of trace elements in the Bolivian Altiplano

Figure 11. Mean trace elements concentrations in soil. a) Aqua regia extraction method (pseudo total concentration) and b) DTPA extraction method (bioavailable concentration), values are on a (mg/kg) dry weight basis.

The results of the study in terms of redox 6.5 Trace element concentrations in sensitive elements in groundwater indicate soils (Paper IV) that Fe shows diverse concentrations in the mining areas in both sites. The concentra- 6.5.1 Mineralogy and soil characteristics tion range of Fe at the Antequera site is 8.5 X-ray diffraction of selected samples in the – 204 Fe µg/L (median, 51.3 µg/L), whereas three (T1, T2 and T3) transects revealed that in Poopó the range is 3.6 – 74 Fe µg/L (me- in T1 the predominant minerals were silicate dian, 14.2 µg/L), except GW-22 (410 µg/L). (quartz), alumino-silicates (muscovite, mer- In thermal water in Antequera (GW-17) the linoite and chlorites). The T2 transect also Fe concentration is 42.8 µg/L and 37.4 µg/L shows silicate (quartz), alumino-silicates in Poopó (Fig. 9a, b; Paper III). (muscovite, albite and kaolinite), while for the T3 transect silicate (quartz), alumino- Second, manganese concentration in the silicates (anorthite, muscovite, and kaolinite) Antequera site exceeds the WHO guideline and magnesium arsenate (Mg2As2O7) consti- value (400 µg/L; WHO, 2011) in four sam- tute the bulk. The presence of latter mineral ples. In Poopó the site, only one exceeds the only occurs in the sample in T3 transect WHO guideline. In thermal water the Mn (Fig. 10; Paper IV). concentration reaches 4.3 µg/L in the Ante- quera site and in Poopó (GW-27 and 28) it 6.5.2 Soil type of studied region reaches a maximum of 270 µg/L (Paper III). The pH of the soil is moderately acid to Thermal water is used for recreational pur- weakly basic in the three sites (total mean poses at both Antequera and Poopó sites, 6.57), mean values were 6.7, 6.4, and 6.61 for but is mixed with surface water (not T1, T2 and T3, respectively. The pH values ARD/AMD) for irrigation purposes. In are homogeneous in the study area, meaning surface water, high concentrations of Al, Fe, that they do not depend on the either crops Mn and Zn are found during both dry and or the soil factors. Soil texture is homogene- rainy seasons, meaning that the same pro- ous among the transects, in the distributions cess governs the release of TEs, the high there are predominantly sand (T1: 27 – 56%, concentrations are results of dissolution of mean 48%; T2: 17 – 56%, mean 46%; T3: 38 abandoned waste, tailing leakage and weath- – 63%, mean 46%), silt (T1: 23 – 44%, mean ering of sulfide minerals (Papers I, II & III). 31%; T2: 20 – 54%, mean 35%; T3: 16 – 36%, mean 30%) and clay (T1: 16 – 28%, mean 22%; T2: 16 – 30%, mean 22%; T3: 17

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

– 25%, mean 22%). The CEC values were 6.5.4 Available trace elements extractions in low (5 – 27 cmol[+]/kg), for T1 range 5.7 – soils 16.8 cmol[+]/kg (mean, 9.3), for T2 5.7 – There are different ranges for available TEs 18.5 cmol[+]/kg (mean, 10.9), and for T3 (Fig. 11b), the range concentrations is 0.03 –

5.8 – 26 cmol[+]/kg (mean, 11.3). The OM 0.35 mg/kg for AsDTPA. The range of As contents were, for T1 0.5 – 2.3% (mean concentration in T1 is 0.04 – 0.12 mg/kg 1.1%), for T2 0.7 – 2.5% (mean 1.3%), and (mean, 0.07); for T2 0.03 – 0.35 mg/kg 0.6 – 2.2% (mean, 1.1%) for T3. The EC (mean 0.09); for T3 0.06 – 0.11 mg/kg mean values in the three transect were 0.12, (mean 0.08). For CdDTPA available concentra- 0.39, 0.34 dS/cm at T1, T2 and T3, respec- tion is 0.02 – 041. The range CdDTPA concen- tively. The electrical conductivity (EC) tration in T1 is 0.03 – 0.11 mg/kg (mean shows that in the T3 the values are higher 0.06), for T2 0.03 – 0.39 mg/kg (mean 0.07), (0.34 dS/cm) than for the others (T1 and for T3 0.02 – 0.41 mg/kg (mean 0.11). For

T2) (Paper IV). CuDTPA the concentration is 0.36 – 1.77. T1 6.5.3 Extraction of major trace element in transect range is 0.58 – 1.21 mg/kg (mean soils 0.91), T2 0.91 – 0.36 mg/kg (mean 0.78), T3 0.64 – 1.77 mg/kg (mean 1.07). For PbDTPA, The Asregia concentrations range from 13 to 38 mg/kg (mean, 23.3) for T1; transect T2 the full range is 0.20 – 1.66 mg/kg, in T1 shows 9.3 – 40.2 mg/kg (mean, 22.1); and 0.48 – 1.66 mg/kg (mean 0.98), T2 0.42 – T3 transect 14.3 – 27 mg/kg (mean, 17.49). 1.64 mg/kg (mean 0.75), and T3 0.20 – 1.04 mg/kg (mean 0.55). The ZnDTPA full range is The Cdregia concentrations in the soils range 0.13 – 36.81, T1 1.16 – 2.45 mg/kg (mean from 0.2 to 2.1 mg/kg, with mean values of 1.67), T2 0.70 – 16.09 mg/kg (mean 4.96), 0.52 mg/kg (range: 0.2 – 0.9) for T1; 0.65 and T3 0.13 – 36.81 mg/kg (mean 7.00). mg/kg (range: 0.3 – 1.2) for T2; and 0.73 The univariate analysis includes the mean mg/kg (range: 0.2 – 2.1) for T3. For Curegia, the concentrations range from 8.66 to 48.58 and range and comparative mean tests were mg/kg, with mean values of 20.52 mg/kg carried out (Student t-test, one-way ANO- (range: 8.7 – 48.6) for T1; 21.04 mg/kg VA) with 0.05 level of significance. The (range: 9.4 – 41.1) for T2, 15.74 mg/kg ANOVA (0.05 level) indicates that there are no significant differences between means at (range: 9.3 – 21.0) for T3. For Pbregia concentrations range from 15.8 to 138.7 mg/kg, the three studied sites for As, Cd, Cu, Pb with mean values of 49.35 mg/kg (range: and Zn (Paper IV). 20.4 – 104.4) for T1; 45.83 mg/kg (range: 6.5.5 Sequential extraction of arsenic in soil 15.8 – 85.9) for T2; and 45.54 mg/kg mg/kg A sequential extraction procedure for As (range: 16.6 – 138.7) for T3. Znregia ranges was applied to selected samples (Wenzel et from 45.54 to 246.61 mg/kg, with mean al., 2001). Exchangeable As (readily availa- values of 92.99 mg/kg (range: 54.5 – 232.5) ble; step 1) ranges from 0.14 to 0.29 mg/kg for T1; 120.18 mg/kg (range: 58 – 203.0) for (mean 0.17) which represents less than 1.9% T2; 96.78 mg/kg (range: 45.5 – 246.6) for T3 (mean 0.8%) of the total As in the samples. (Fig. 11a; Paper IV). This fraction is indicating the most im- The univariate analysis includes the mean portant As pool related to environmental and range and comparative mean tests were risk (Wenzel et al., 2001). The T3 transect carried out (Student t-test, one-way ANO- samples have slightly higher contents than VA) with 0.05 level of significance. ANOVA the T1 and T2 samples. test of extracts of aqua regia TEs (As, Cd, Cu, Pb and Zn) reveal that there is no significant difference among the three places (p> 0.05, one-way) indicating their probably common source.

Geochemistry of trace elements in the Bolivian Altiplano

Figure 12. Trace elements (heavy metal) fractionation ( F 1 – F4) in sediments of Lake Titi- caca.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

For specifically sorbed As (step 2), the full that in both sites the content of OM varies range constituted 0.60 – 2.84 mg/kg (mean between 1.5 – 5.6% and 6.1 – 13.2% in shal- 1.13). The As released in this fraction can be low and deeper areas, respectively. The re- potentially mobilized and directly take up by coveries are in the range 90 – 105%, and the plants due to competition between phos- total extraction and sum of the fractions phate and arsenate for adsorption sites indicate good agreement. The total TE con- (Wenzel et al., 2001; Niazi et al., 2011). As centrations vary as follows: Cu (6.7 – 55.5 associated to amorphous Fe oxides (step 3) mg/kg), Zn (14.9 – 940.9 mg/kg), Pb (29.7 represents the largest pool with the range – 161.2 mg/kg), Cd (0.3 – 3.2 mg/kg), Fe from 4.29 – 12.06 mg/kg (mean 7.14). As (883 – 5941 mg/kg), Mn (83.3 – 622.1 extracted from crystalline Fe oxides (step 4) mg/kg), Co (14.7 – 27.7 mg/kg) and Ni (6.1 represents the range 2.30 – 9.96 mg/kg – 20.6 mg/kg). (mean 6.05). The residual phase (step 5) has The TE concentrations show that Cd, Cu a general range from 1.68 to 28.35 mg/kg and Zn were found from very low to mod- (mean 7.62) (Paper IV). erate levels at all sampling locations. There is 6.6 Trace element concentrations in a moderate contamination by Pb for all crops sampling points except for S4, it coincides The data for dry weight basis were convert- with a high contamination level by Pb and ed assuming a mean water content of 90% Zn; this was the place for transfer the min- (Rötting et al., 2013) in order to compare erals, loaded from train to boat and trans- with Chilean and international regulations ported to Puno Harbor (Peru). Also, Ni con- (FAO, WHO, EC). tamination for all points appears to be very high in some fractions. The As concentration in beans ranges from Sequential extraction (Fig. 12) results show 0.19 to 0.24 mg/kgfw (mean 0.19); for Cd the range is 0.02 – 0.05 mg/kg (mean 0.03); that TEs associated to the exchangeable and fw acetic acid soluble fraction (F1) decrease in and for Pb 0.077 – 0.499 mg/kgfw (mean 0.319). The concentrations in barley is 0.093 the order Fe> Mn> Zn>Pb >Co >Ni> Cu> Cd. The TEs associated with the reduc- – 0.396 mg/kgfw (mean 0.172) for As, 0.008 – 0.029 mg/kg (mean 0.012) for Cd, and ible fraction (i.e. bound to Fe and Mn ox- fw ides), fraction 2 (F2), decrease in the order 0.038 – 0.401 mg/kgfw (mean 0.162) for Pb. For potato the As concentration is 0.093 – Fe> Zn> Mn> Pb> Cu> Ni> Co> Cd. 0.208 mg/kg (mean 0.122), Cd 0.008 – Metals associated with OM and sulfides, fw fraction 3 (F3), decrease in the order Fe> 0.044 mg/kgdw (mean 0.023), and Pb 0.038 – 0.493 mg/kg (mean 0.156). Mn> Zn> Pb> Cu> Co> Ni> Cd. The fw metals associated with the residual fraction The univariate analysis includes the mean, (i.e. strongly linked with the crystalline struc- median, range and variance; comparative ture of minerals) fraction 4 (F4), decrease in mean tests were carried out (Student t-test, the order Fe> Mn> Zn> Pb> Cu> Co> one-way ANOVA) with 0.05 level of signifi- Ni> Cd (Paper V). cance. Results for As and Pb show that there are no significant differences between mean 7 DISCUSSION (p<0.05; one-way) in the four different crops; while Student´s t-test (0.05 level) for The composition of natural waters is closely related to the composition of the rocks that Cd shows that there is a significant differ- they contact, transform and partially dissolve ence in means between beans – barley and (Morel, 1983). barley – potato crops (Paper IV).

6.7 Trace element concentrations in lake sediments (Paper V) Nine sediments samples at different depths (8 – 156 m) of Minor and Major Lakes show 28

Geochemistry of trace elements in the Bolivian Altiplano

7.1 Mechanisms controlling the hydrochemistry in the Poopó Basin (PB) (Paper I) The natural water composition is controlled by both geochemical and geo-biological processes (not discussed here) (Morel, 1983). The groundwater composition depends on the bulk mineralogy of the aquifer sediment as well as on chemical weathering, one of the major processes controlling the geochemical cycle of elements (Morel, 1983). During this process, rock and primary minerals become transformed to solutes and secondary minerals (Stumm & Morgan, 1996). The pH of water is then primarily determined by a balance between dissolution of weakly acidic carbon dioxide

(CO2) and basic rocks, alumino-silicates and carbonates (Morel, 1983). Taking this into account, the geochemical evolution of the groundwater around Poopó Basin (PB) depends on whether the water source is near to Eastern Cordillera or in the alluvial fan of the PB. For the whole basin - the major anions follow the order HCO3 > - 2- + Cl > SO4 and the major cations are Na > Ca2+ > Mg2+ (Paper I). The existence of the

HCO3–Na water type can be due to cation exchange, as reported by other studies (Bundschuh et al., 2004; Ravenscroft & McArthur, 2004; Bhattacharya et al., 2006a). Figure 13. Box and whisker plots show- This also agrees with the calcite, kaolinite ing the distribution of all the major ions clay mineral, dolomite and gypsum satura- in a) groundwater and b) surface water. tion indices (SI) calculated by PHREEQC,

Figure 14. Bivariate plots between a) Na+ versus Cl- and b) Ca2+ + Mg 2+ versus SO42- + HCO3- in groundwater samples.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Table 1. Ionic ratios of groundwater and surface water samples from the Antequera and Poopó sub-basins (Ratio derived from meq values). Ionic ratio Antequera GW Poopó GW Antequera SW Poopó SW min max mean min max mean min max mean min max mean Ca2++ Mg2+/Na++ K+ 0.1 4.2 1.3 0.1 5.0 1.6 0.4 30.5 10.9 0.3 4.9 2.1 HCO3-/(Ca2++ Mg2+) 0.0 1.6 0.4 0.0 0.7 0.2 0.0 0.0 0.0 0.0 0.8 0.1 - 2+ HCO3 / Ca 0.0 2.4 0.5 0.0 0.8 0.32 0.0 0.01 0.0 0.0 1.03 0.16 Ca2+/Na+ 0.1 4.2 1.2 0.1 4.3 1.5 0.4 34.2 11.4 0.2 4.1 2.0 Mg2+/Na+ 0.0 1.1 0.3 0.0 1.0 0.3 0.1 1.5 0.9 0.0 2.0 0.8 Ca2+/Mg2+ 2.2 8.5 5.5 2.2 12.0 5.8 2.6 52.3 12.0 1.6 12.4 3.8 2+ 2+ Ca /SO4 0.5 9.4 2.4 0.8 5.5 2.4 0.7 6.2 1.7 0.6 3.4 1.5 Mg2+/Ca2+ 0.1 0.5 0.2 0.1 0.5 0.2 0.0 0.4 0.2 0.1 0.6 0.4 Cl-/Na+ 0.4 1.5 0.9 0.6 1.9 1.1 0.6 3.7 1.7 0.6 1.4 1.0 - - 2- HCO3 /HCO3 + SO4 0.0 0.9 0.3 0.0 0.7 0.3 0.0 0.0 0.0 0.0 0.6 0.1 Ca2++ Mg2+ 0.9 26.9 10.7 4.2 19.2 9.1 2.2 68.6 25.8 1.6 93.2 16.1 as also reported by other authors (Banks et spectively (Table 1). For both sub-basins al., 2004; Lilja & Linde, 2006; Selander & there seems to be a limited enrichment from Svan, 2007) (Fig. 13a). the dissolution of silicate minerals. Weather- + - ing of carbonates does not seem to be a pro- In the entire PB, Na versus Cl are strongly 2 cess controlling the hydrochemistry as indi- positively correlated (r = 0.8; Fig 14a) (Paper cated by Na+ + K+ versus HCO - and be- I), indicating the presence of soluble halite 3 cause the mean values of the HCO -/(Ca2+ + (NaCl) in the groundwater in the study area, 3 Mg2+) ratio are relatively low (0.4 and 0.2 for due to i) high evapotranspiration rate, ii) precipitation of NaCl during the dry periods the Antequera and Poopó sub-basins, respectively; Fig. 15a, b). The bivariate plots and iii) dissolution of salty Quaternary sedi- show two distinct trends, one along the y = ments (Minchin, 1882; Coudrain-Ribstein et al., 1995a, b; Argollo & Mourguiart, 2000; 2x (Fig. 15a), which suggests that carbonate mineral dissolution is limited in some sam- Fornari et al., 2001; Zapata, 2011). The plot - 2- 2+ 2+ ples, whereas the second trend indicates that HCO3 + SO4 versus Ca + Mg is less 2 there may be a combination of carbonate correlated (r = 0.7, Fig. 14b; Paper I) sug- and silicate dissolution in so much as sam- gesting that dissolution of calcite, dolomite and gypsum are reactions occurring in the ples are located between the y = 2x and y= x. system as reported by other authors in the 2+ 2+ PB (Banks et al., 2004; Lilja & Linde, 2006; The scattered plot between Ca versus Mg 2+ - + Selander & Svan, 2007). There are additional and Ca versus HCO3 (both axes Na nor- sources of sulfate such as weathering of sul- malized; Fig. 16a, b) suggests that although fide minerals due to the geological condi- many groundwater samples in the zone are tions of the study area. indicative of global silicate weathering, other samples in the zone indicate evaporite min- 7.2 Distinguishing hydrochemistry eral dissolution, whereas some samples be- mechanisms in the Antequera and tween these zones and indicate the occur- Poopó Sub-basins (Paper II) rence of both processes. 7.2.1 Discrimination between processes related The cation exchange can be evaluated by to groundwater chemistry + - 2+ 2+ bivariate plots (Na - Cl ) versus (Ca + In groundwater samples, mean (Ca + 2+ - 2- 2+ + + Mg )-(HCO3 + SO4 ) (Fig. 17) where the Mg )/(Na + K ) ratios in the Antequera slope should be -1 (i.e. y = -x) (Jankowski et and Poopó sub-basins were 1.3 and 1.6, real., 1998; Biswas et al., 2012). Hence, the 30

Geochemistry of trace elements in the Bolivian Altiplano

Figure 15. Bivariate plots for a) (Ca2+ + Mg 2+) versus HCO3-, and b) (Na+ + K+) versus - HCO3 in the Antequera and Poopó sub-basins. Open diamonds, groundwater samples, and solid triangles, surface water samples. slope values of -0.76 and -0.71 in the Ante- enrichment of K+ in the Antequera ground- quera and Poopó sub-basins, respectively, water (Mukherjee et al., 2009). indicate that cation exchange is to some ex- Scatter plots are shown in Figures 18 and 19. tent responsible for the composition of shal- There is a strong correlation between Na+ low groundwater. The higher correlation in - 2 2 and Cl in both groundwater (r = 0.99; Fig. the (r = -0.76) Antequera groundwa ter 2 2 18a, b) and surface water (r = 0.97; Fig. 19a) samples in comparison to the (r = -0.71) samples. This probably indicates both evap- Poopó samples also suggest that cation oration and the dissolution of paleolake salts exchange might contribute to the relative

Figure 16. Bivariate plots for a) Ca2+/Na+ versus Mg2+/Na+, and b) Ca2+/Na+ versus HCO3-/Na+ in the Antequera and Poopó sub-basins. Open diamonds; groundwater samples and solid triangles; surface water samples.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Figure 17. Bivariate plots of HCO3- + SO42- corrected Ca2+ + Mg 2+ versus Cl- corrected Na+. Solid diamonds, Ante- quera groundwater samples; open diamonds Poopó groundwater samples and solid squares, Antequera and Poopó surface water samples. such as halite, as discussed by Coudrain- Ribstein et al. (1995a, 1995b), Argollo and Mourguiart (2000) and Fornari et al. (2001). In some groundwater samples the ratio between Cl-/Na+ exceeds 1 (Table 1), and indicates that the enrichment of Cl- in shallow groundwater relates to anthropogenic sources such as agricultural residues (Jack et al., 1999) and to the atmospheric deposition of previously re-suspended soil dust particles Figure 18. Bivariate plots for groundwa- originating from ore deposits (Goix et ter showing a) Na+ versus Cl- (large al., 2012). scale), b) Na+ versus Cl- (small scale), - 2- 2+ - 2- 2+ 2+ and c) HCO3 + SO4 versus Ca + The plots HCO3 + SO4 versus Ca + Mg Mg2+. Solid diamonds, Antequera sam- (Fig. 18c) in groundwater show poor correla- ples; open diamonds, Poopó samples. tion, suggesting that calcite, dolomite and gypsum dissolution is limited in both sub- to the high concentration of SO 2- and Zn. basins. In contrast, surface waters in both 4 However, this explanation is not consistent sub-basins show a strong correlation be- 2- - 2- 2+ 2+ with the near-neutral pH values. High SO4 tween HCO3 + SO4 versus Ca + Mg 2+ 2 2 and Ca concentrations at these sampling (Fig. 19b; r = 0.86 Antequera, r = 0.99 Po- sites may also be due to gypsum dissolution. opó), which indicates that the dissolution of Thus, it appears that SO 2- concentrations calcite, dolomite and/or gypsum can be de- 4 are influenced by the reactions in Eqs (3) scribed by the following equations (Sun et and (4), which represent water-rock reac- al., 2013). tions that can take place due not only to the 2+ CaMg(CO3)2(s) + 2H2O + 2CO2 → Ca + attack of CO2 dissolved in the water, but 2+ - Mg + HCO3 (1) also by the attack of H2SO4 produced by 2+ - oxidation of sulfides (Sun et al., 2013). CaCO3(s)+ H2O + CO2 → Ca + 2HCO3

(2) 3CaxMg1-xCO3(s) + H2CO3 + H2SO4 → 2+ 2+ - 2- The oxidation of sulfide minerals is suggest- 3xCa + 3(1-x)Mg + HCO3 + SO4 (3) 2+ 2- + ed by data from GW-1 – GW-4 and GW-11 FeS2(s) + 7/2O2 +H2O → Fe + SO4 + 2H at the outlet of the Antequera sub-basin due (4)

Geochemistry of trace elements in the Bolivian Altiplano

are predominant Ca2+ > Na+ > Mg2+, whereas Na+ > Ca2+ > Mg2+ are predominant in Poopó surface waters. The principal anions 2- - 2- - are SO4 > Cl > CO3 > HCO3 in Ante- 2- - 2- - quera and SO4 > HCO3 ≈ CO3 > Cl in Poopó surface waters (Paper II). Median 2+ 2- concentrations of Ca and SO4 were much higher in Antequera than in the Poopó surface water samples, indicating that gypsum dissolution may play a major role in the Antequera sub-basin. During both dry and rainy seasons along the main stream (Ante- quera River) it is perceived that the predominant process is the oxidation of the wastes abandoned on the banks of the river. In surface water samples, mean (Ca2+ + Mg2+)/(Na+ + K+) ratios in both Antequera and Poopó sub-basins are 10.9 and 2.1, respectively (Table 1). Antequera samples show a clear signal of dissolution of silicate minerals, especially in waters affected by AMR/AMD. Samples in the headwaters that are not affected by mining activities show high (Ca2+ + Mg2+)/(Na+ + K+) ratios (1.8 – 4.6), which indicate that dissolution of carbonate minerals exerts a dominant control on the local hydrochemistry (Table 1).The scattered plot between Ca2+ versus Mg2+ and 2+ - + Ca versus HCO3 (axes Na normalized; Fig. 16) shows that surface water samples Figure 19. Bivariate plots for surface indicate silicate weathering dissolution and water showing a) Na+ versus Cl-, and b) little dissolution of evaporites. Samples af- 2- HCO3- + SO42- versus Ca2+ + Mg 2+. Sol- fected by AMD and ARD show high SO4 - 2- id squares, Antequera samples; open /Cl ratios (range 3 to 18), i.e. are SO4 en- squares, Poopó samples. riched (Table 1). Most surface water samples from the Ante- 7.2.2 Describing the processes in surface quera and Poopó sub-basins plot in the re- water chemistry gion dominated by rock weathering (indicat- During the dry season, the surface water ed by a circle, Fig. 20). In the evaporation- chemistry in the Antequera River in areas crystallization dominated region of the impacted by mining activities has a particular Gibbs ratio plot (Fig. 20, indicated by an behavior, where almost all samples have asterix), sample SW-6 from the Antequera 2- 2+ sub-basin is affected by ARD and AMD, SO4 and Ca as dominant ions at R-II (Pa- 2- and sample SW-18 from the Poopó sub- per I). The elevated SO4 concentrations in the surface water reflect oxidation of the basin comes from the tailings area. The re- mining wastes, consisting mainly of sulfide maining three Poopó and two Antequera minerals such as pyrite, and evaporation samples plot between the rock weathering enrichment around the eastern shore of and evaporation-crystallization domains. Lake Poopó. During the rainy season, in the main stream of the Antequera river the following cations

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Figure 20. Gibbs ratios to indicate processes controlling surface water chemistry. Solid squares; Antequera samples, open squares Poopó samples. Other values indicate (1) Mean US precipitation, (2) Negro River, (3) Orinoco River, (4) Ganges River, (5) Nile River, (6) Volga River, (7) Rio Grande, (8) Colo- rado River, (9) Jordan River and (10) World oceans (Gibbs, 1970).

7.3 Mobilization of trace elements adsorption on ferrihydrite surfaces in a part and As in the hydrologic system of the study area. The groundwater with Dissolution and formation of solid hydrox- high As concentration in some sites of re- ides and oxides are important in aquatic gion II also reflects localized patches where chemistry of many metal ions, particularly Fe the process of desorption from Fe(III) ox- and Mn that depend on pH (Morel, 1983). ides and hydroxides is possibly one of the But also other ligands such as carbonate, principal mechanisms controlling the mobili- sulfide or phosphate, whose free concentra- ty of As, due to moderately oxidizing conditions depend not only on pH, but also on tions (Smedley & Kinniburgh, 2002). The total complex concentration and concentra- hydrous ferric oxide (HFO) adsorption on tions of precipitating metals (Morel, 1983). surfaces is supported by modeling results that reveal that the groundwater was over- 7.3.1 Mobilization of trace elements and As saturated for Fe(III) phases such as ferrihy- in the groundwater environment in the drite. In contrast, the groundwater was un- Poopó Basin (PB) (Paper I) dersaturated with respect to all As minerals. In PB, there is a considerable variability in Thus, the poor correlation of As versus TEs groundwater chemistry, redox conditions and As versus SO 2-, has led to the conclu- and TEs and As, the groundwater shows 4 sion that the distribution of As and TEs is two different trends: first, samples with high not well defined, but seems to be controlled As concentrations (100-250 µg/L) in some by surficial oxidation processes. wells in both regions III and V, and second, samples with low concentrations of As (up There are three different types of impact in to 55 µg/L) in wells in regions I, II, and IV sub-basin and transects, which are summa- (Fig. 8). rized as: i) natural by geological fingerprint (Tapia et al., 2012), ii) mining activities and In general, the low As and Fe concentrations tailings (PPO-04, 1996), and iii) clearing and in the groundwater suggest that As distribu- debris uncovered (PPO-04, 1996). tion is controlled by the co-precipitation and

Geochemistry of trace elements in the Bolivian Altiplano

7.3.2 Mobilization of trace elements and As sources of contamination, which sometimes in groundwater in the Antequera and overlap each other. The seasonal variation in Poopó Sub-basins (Paper III) the transportation of TEs by re-suspension Consequently, samples collected from the of soil dust and dry deposition (Goix et study area indicate almost the same chemical al., 2011), on soils by wind has not been variability and to interpreted geochemical evaluated in this work. mechanisms is complicated by the different It is worth mentioning that TE and As mobilization will be discussed in this section. The factor analysis approach was used to better understand the relationship between parameters. In groundwater in the Ante- quera Sub-basin, five factors explained most of the total variance in the geochemical data set (86.7% of the variance; Paper III) from groundwater samples (Fig. 21 and 22). Trace elements show a positive loading in all the cases for both factors 2 (20%) and 3 (14%) which suggests the weathering of minerals (plagioclase, evaporates, calcites, and gypsum) as controlling the release of major TEs (Al, Fe, and Mn); the negative correlation of Al and Fe with respect to pH suggests the pH dependency of mineral solubility. The negative correlation between Eh and pH would seem to be due to the thermodynamic stability of water, and the positive loading 2- 2+ 2+ for SO4 , Ca , Mg and Mn, may represent gypsum dissolution (Coudrain-Ribstein et al., 1995a, 1995b). Factor 4 (10%): the positive loading for Cd and Zn may indicate remobilization of TEs from hydrous ferric oxide (HFO) surfaces. Cd is more chalcophile than Zn and tends to persist in a sulfide while ZnS is weathering (Fuge et al., 1993) which explains its relative immobility in surface soils in the present study (Paper IV), as was

observed in the CdDTPA and ZnDTPA extracts. Factor 5 is probably related to competing F i g u re 21. Eh – pH diagram in a) whole 3- adsorption sites between As and PO4 Poopó Basin and b) Antequera and Poopó (Signes-Pastor et al., 2007). This finding is sub-basins. Red numbers, groundwater also supported by the As speciation in soils, samples, and black numbers, surface water for As specifically sorbed; it is the fraction samples. Green line show the evolution of As(V) species in surface water samples potentially mobilized as a result of competi- along Coro Coro river from headwater to tion between phosphate and arsenate ad- outlet sub-basin; blue line show the evolu- sorption sites, due to changes in pH or tion of As(V) species from Bolivar tailing phosphate application to soils (Wenzel et al., to Avicaya sector before to mix with Coro 2001; Niazi et al., 2011). Coro rivers until sub-basin outlet.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Figure 22. Factor loadings for different geochemical parameters in Antequera groundwater. Load- ing values ≥±0.5 represent a significant contribution of the corresponding parameter (Paper III).

Figure 23. Factor loadings for different geochemical parameters in Poopó groundwater. Loading values ≥±0.5 represent a significant contribution of the corresponding parameter (Paper III).

Geochemistry of trace elements in the Bolivian Altiplano

In Poopó groundwater samples (Fig. 21 & and 2 make it possible to distinguish be- 23), four factors explained most of the total tween dissolution processes in addition to variance in the geochemical data (83.9%; TE mobilization in different areas of the Paper III). TEs show a positive loading in all Antequera sub-basin. factors. Factor 1 (42%) shows a positive Four factors also explain most of the vari- loading for EC, Cl-, SO 2-, Al, K+, Mg2+, Na+ 4 ance (96.0%) in the Poopó surface waters and Si, suggesting the weathering of minerals (Fig. 21 and 25). Factor 1 shows a positive (plagioclase, evaporates and carbonates). loading for TEs (Al, As, Cd, Fe, and Zn). Factor 2 (18%), has positive loadings for Eh, 2+ These could be linked to three different geo- alkalinity, Ca , and Fe and negative loadings chemical processes: i) the weathering of pla- for pH and Cd, which may represent the gioclase and dissolution of halite and gyp- dissolution of calcite and consequently in- sum, ii) the mobilization of TEs due to the creasing the Fe solubility at low pH and also dissolution of minerals at acidic pH (most the precipitation of Cd (CdCO ) (Rötting et 3 waters were probably later mixed with more al., 2006) with enrichment of this TE in the alkaline waters, therefore explaining the lack solid phase. Factor 3 (14%) has positive of correlation with pH), and iii) the dissolu- loadings for alkalinity, As, Cd, Mn and Si, tion of gypsum and halite. Factor 4 (6%) showing first that Cd can be inmobilized by shows a low positive loading for EC and precipitation as carbonate and second sug- Mn. However, Factors 3 and 4 only explain a gesting that dissolved HCO - and Si may act 3 very small fraction of the total variance and as competitors for As at the adsorption sites can therefore be considered as having only on Mn and Fe hydroxides (Anawar et al., little significance in the geochemistry in the 2004), thereby increasing the As dissolved in Poopó sub-basin (Paper III). groundwater. Finally, factor 4 (10%) indicates the dissolution of gibbsite and there- 7.4 Mobilization of trace elements fore the mobilization of TEs (Al and Zn). and As in transect soils 7.3.3 Mobilization of trace elements and As (Paper IV) in surface water (Paper III) 7.4.1 Mineralogical description of soils On the main stream samples from the Ante- The X-ray diffraction patterns are similar quera sub-basin, factor analysis was carried along the three locations, quartz, muscovite, out and four factors explained 91.5% of the and kaolinite constitute the bulk mineral in variance in the data set (Fig. 21 and 24). Fac- the three transects. But there are also other tor 1 (51%) has positive loadings for EC, minerals such as merlinoite and chlorites in 2- 2+ Eh, SO4 , Al, Cd, Mg , Si and Zn, and neg- T1, in T2 calcium-albite and in T3 anorthite ative loadings for pH and alkalinity. This and magnesium arsenate (Mg2As2O7). The suggests three geochemical processes: i) the latter is identified by X-ray diffraction as an weathering of plagioclase, ii) the mobiliza- intensive signal in position 2Θ at 28.54, its tion of TEs due to the dissolution of miner- presence in this sample and no other sam- als at acidic pH, and iii) the oxidation of ples is remarkable (Fig. 10). The occurrence sulfide minerals. Factor 2 (17%) has positive of kaolinite in inundated alluvial plains indi- 2- 2+ loadings for EC, SO4 , Ca , Cd, and Mn, cates the ultimate chemical alteration of the and negative loadings for pH, alkalinity and predominant silicate rocks present in the 3- PO4 . This can be linked to the dissolution three sites (Fig. 10). The resolution of the X- of gypsum, both factors are linked to weath- ray diffraction and the sensitivity of the radi- ering of plagioclase, gypsum and evaporite ation source (CuKα) used in the current (Coudrain-Ribstein et al., 1995a, 1995b) study was not suitable for identification of minerals at acidic pH. Factor 4 shows posi- other Fe-hydroxide minerals. Comparative tive loadings for As and Fe, which is linked analysis of the three transects shows no vari- to the adsorption of As onto Fe hydroxides ations which is also true for the profile sam- and to As and Fe mobilization when these ples in the Poopó alluvial fan (POM2, 4, 7 precipitates re-dissolve. Therefore, factors 1 and 9 m deep) (Paper IV) 37

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Figure 24. Factor loadings for different geochemical parameters in Antequera surface water. Loading values ≥±0.5 represent a significant contribution of the corresponding parameter.

Figure 25. Factor loadings for different geochemical parameters in Poo- pó surface water. Loading values ≥±0.5 represent a significant contribution of the corresponding parameter.

Geochemistry of trace elements in the Bolivian Altiplano

7.4.2 Characteristics of trace elements in carbonate is not dominant, this matches the soils: trends with aqua regia extractions results of step1 (As fractionation) In the current work, the contents of TEs in (Paper IV). aqua regia extracts show that As, Cd, Cu and 7.4.4 Characteristics of arsenic in soils: Pb are lower than the median values report- trends with sequential extractions ed by PPO (1993-1996), only the Zn con- Arsenic easily released from non-specifically tents are slightly higher than those in the adsorbed sites in the soils can be found in PPO (1993-1996) report. The present study fraction 1, for As associated to carbonates and the PPO data base are directly compa- and according the results of Pearson´s and rable due to use of a similar extraction Spearman´s correlation their presence is not method. In comparison with world soils predominant in the solid phase in the study database, contents of As, Cd, Pb and Zn are area, but this fraction represents the As con- higher, while the concentrations of Cu is centrations (< 0.29 mg/kg; mean 0.17) that lower than the average values for world soils can be used for predicting solute As, and is (Han, 2007), suggesting that the three areas useful for risk assessment of As leaching to have high background content of TEs, the aqueous phase (Wenzel et al., 2001). The probably related to geologic conditions in relatively high content in the T3 transect can the Eastern Cordillera with the presence of a be supported by the XRD soil results, it can polymetallic (Ag-Pb-Zn-Sn-Sb) belt in this be a source of As from dissolution of mining district (Arce-Burgoa & Goldfarb, Mg As O mineral (Paper IV). 2009; Tapia et al., 2012) (Paper IV). 2 2 7 The Asstep-2 specifically sorbed can be poten- Pearson´s correlation is slightly positive tially mobilized due to pH and phosphate 2> 0.43) among aqua regia extracts of TEs (r concentration in soil and according to com- (As-Cd, Cu, Pb, Zn), (Cd-Pb, Zn), (Cu-Pb, petition between phosphate and arsenate for Zn), (Pb-Zn), it could indicate a similar geo- adsorption sites discussed in paper III (fac- chemical behavior and/or common source tor 5), elevated phosphate concentrations in material of these elements (Burak et al., soil effectively de-sorb As(V), increasing 2010) (Paper IV). The correlation between dissolved As concentrations (Signes-Pastor Fe/Mn-As, Cu, Pb and Zn (p< 0.01) seems et al., 2007) (reactions #12.1, #12.2; Paper to be related to the formation of secondary III). The soluble phosphate in soil ranges iron oxides with capacity to strongly adsorb from 5.95 to 15.54 mg/kg in the T3 transect these TEs, as already described in the litera- (Paper IV). ture (Burak et al., 2010). The last three fractions constitute the largest 7.4.3 Characteristics of trace elements in pool of As (~ 90%), suggesting that As as- soil: trends with DTPA single sociated to amorphous Fe oxides is the main extraction source (F3= F5> F4; Paper IV) of As re- Significant positive Pearson´s and Spear- lease due to changes of pH, redox, OM and man´s rank correlations are found among environmental conditions (minerals and the available TEs, indicating that these TEs drainage). are potentially available for plants (Hong et al., 2008). Meanwhile comparing with physi- 7.4.5 Mobilization of trace elements in lake cal-chemical characteristics only the OM- sediments

AsDTPA shows a correlation significant at the The sedimentation dynamics are still poorly p< 0.01 level (Spearman’s rank), suggesting known, in the case of carbonate deposits that this As could be linked to OM. Fur- Boulangé et al. (1981) give a rate of 0.5 thermore, carbonate species are also estab- mm/year for Lake Major and 10 times high- lished as potentially bioavailable, but the er for Lake Minor; the rates are controlled slightly negative (r2= -0.51; p< 0.01) Spear- by the relation between allochthonous ele- man´s rank correlation between soil pH- ments of detrital origin and autochthonous

AsDTPA suggests that the retention of TEs as

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04 elements of biogeochemical origin (Rodrigo Cu bound to OM is well-known in geo- & Wirrmann, 1992). chemistry (Pempkowiase et al., 1999; Gon- Both deep and shallow sediment of Lake záles et al., 2009), which under oxic condi- Titicaca has shown varying TE spatial distri- tions can be remobilized to the water col- bution based on the geological evolution and umn. The riparian towns are the main source anthropogenic conditions around the lake. of domestic residues in S3, S4, S5 and S6. However, the low content in S7, S8 and S9 Earlier studies by Iltis et al. (1992) revealed a in F3 seems to be due to slightly higher pH predominantly anoxic environment in the values and photosynthetic activities (Iltis et sediments of both the Lake Major and the al., 1992) taking up Cu, which explains the Chua depression in the Lake Minor, which lower Cu available in fraction F1 (< 4.16 linked to OM in sediments that can cause mg/kg). the mobilization of species that otherwise would be retained by adsorption due to high The behavior of Ni is highlighted due to the CEC values (Paper V). fact that almost all Ni is extracted in the first three fractions, which can be easily remobi- There is a greater content of higher heavy lized by changes of environmental condi- metals in the Lake Major than in the Lake tions (pH, redox potential, dissolved oxygen, Minor. This is consistent with the fact that etc.). Nickel occurrence in the sediments Lake Major receives on average nearly 85% follows the order: F3> F2> F4> F1 of the total inputs of dissolved salt com- (Paper V). pared to only 15% for Lake Minor (Car- mouze et al., 1992). The total dissolved salt Zn partitioning gives the order F1= F2> concentration varies from 0.9 g/L in Lake F4> F3 (Paper V). The same content is Chucuito (Lake Major) to 1.2 g/L in Lake found in the F1 and F2 fractions; in con- Huiñaimarca (Lake Minor), meaning that the trast, it is high in samples S1, S4 and S5 due - 2- + to mining activities in these representative waters are dominated by Cl , SO and Na 4 places (Passos et al., 2010). Zinc may be co- (Carmouze et al., 1981; 1992). precipitated with isostructural carbonates,

The behavior of the TEs in the lake sedi- e.g. siderite (FeCO3). Additionally, the con- ments is special. The Fe partitioning (Paper tent of Zn in oxides/hydroxides is con- V) in the sediments follows the order: F4> sistent with the observed chemical fractiona- F3> F2> F1. S1 and S4 samples show high- tion of Zn in F2 (Ettler et al., 2006). Again, er content in F4, but other samples show S7, S8 and S9 show low content; it could be almost the same relative low content, sug- explained by high macrophyte activity that gesting that this could be considered a back- takes up Zn. ground value (Rodrigo & Wirrmann, 1992). The lower content of Fe in the F3 fraction is Lead is mostly associated to the residual and related to great photosynthetic activity of oxidizable fraction (F3> F4> F2> F1; Paper V) and Cd fractionation indicates that the phytoplankton and abundant macrophytes main portion is associated to OM phases (Iltis et al., 1992) in samples S7, S8 and S9. The Mn partitioning in the sediments fol- (S2, S4, S7, S8 and S9; F3> F2= F1> F4). lows the completely opposite order to the Cd can adsorb onto Fe and Mn oxides that distribution of Fe: F1> F2> F3 >F4 (Paper are found in S4 in fraction F2; meanwhile F1 V). It is possible that the formation of (exchangeable and carbonate fraction) is related to shallow sediments with high mac- MnCO (Karageorgis et al., 2009) occurs in 3 rophyte activities (González et al., 2009). the (S1, S5 and S6) samples related to the F1 fraction. In contrast, the mineralization of The partitioning of Co in the sediments fol- fraction F3 could easily release Mn to the lows the order F1> F3> F4> F2. The re- water column in samples S3 and S4. mobilization of Co from sediment to water occurs under anoxic conditions and stops The Cu partitioning in the sediments follows when the water becomes oxic (Petersen et the order F3> F4> F2> F1 (Paper V). The al., 1997) (Paper V). 40

Geochemistry of trace elements in the Bolivian Altiplano

7.5 Arsenic and other trace elements broad beans are baked with the shell (Quei- in crops rolo et al., 2000). Among the countries in Latin America few Cadmium have legislation and health considerations The range for Cd in different crops varied concerning TEs in food; many countries from 0.0002 to 0.040 mg/kgfw in Chile have no legislation on these points, and nei- (Queirolo et al. 2000). In potato the Cd ther does Bolivia, so the comparison will be mean concentrations is 0.0094 mg/kgfw, and made with international regulation (FAO, broad beans show 0.0033 mg/kgfw (Queirolo WHO, EC). The discussion in this section is et al., 2000). Oporto et al. (2007) reported a limited to three TEs (i.e. As, Cd and Pb) due Cd concentration of 0.21 mg/kgfw in pota- to their toxicity to human and also to plants toes in contaminated soils in the agricultural (McBride, 1994; Kabata-Pendias & Mukher- valley at Pilcomayo River, where 81% of the jee, 2007). samples exceed the international regulations. Arsenic In this study, Cd mean contents are: beans Queirolo et al. (2000) concluded that pota- (0.03 mg/kgfw) > potato (0.023 mg/kgfw) > toes (including skin) constitute an efficient alfalfa (0.017) > barley (0.012 mg/kgfw). The bioaccumulator of As with a maximum con- mean global Cd content is 0.021 mg/kgfw, not exceeding the international (FAO, centration of 0.86 mg/kgfw, while beans showed a maximum concentration of 0.112 WHO, EC) regulation guidelines which es- tablish limits for potatoes (peeled) and mg/kgfw. Muñoz et al. (2002) determined that concentrations of As in the skin was stem/root vegetables (0.10 mg/kg, fresh between 3 to 70 times greater than the levels weight) and other vegetables (0.05 mg/kg, in the edible part, for potato the ratio is 3 fresh weight) (JECFA, 2005; European times, the concentration in the edible part is Commission Regulation (EC-1881), 2006). 0.021 mg/kgfw. Comparison with the con- Lead tents found in the edible part in the present The range for Pb in different crops varied study shows the following concentration of from 0.0006 to 0.094 mg/kgfw (Queirolo et As to be: beans and alfalfa (0.19 mg/kgfw) > al., 2000). The Pb total mean concentration barley (0.17 mg/kgfw) > potato (0.11 in the crops is: for alfalfa 0.025 mg/kgfw, for mg/kgfw). The total mean (i.e. taking into potato 0.011 mg/kgfw , for broad beans 0.023 account all the crops analyzed in this study) mg/kgfw. (Queirolo et al., 2000). In this As content is 0.16 mg/kgfw; no samples ex- study, high Pb contents are found in beans ceed the tolerable limits for crops. The Chil- (0.32 mg/kgfw) > alfalfa (0.22 mg/kgfw) > ean regulation for food (Muñoz et al., 2002) barley (0.16 mg/kgfw) ≈ potato (0.16 establishes 0.5 mg/kgfw, for cereals and leg- mg/kgfw). The European regulation guide- umes, and 1.0 mg/kgfw for other products. lines have a maximum level of 0.20 mg/kg The contents measured here are fivefold (fresh weight) for cereal and legumes (barley, below the regulations. Also, these As con- alfalfa and beans), and 0.10 mg/kg for tents are lower than those found in a review peeled potatoes (European Commission database (mean: 6.22 mg/kgfw; range: 0.005 – Regulation (EC-1881), 2006). The beans, 56.9 mg/kgfw) reported by Rötting et al. alfalfa and potato exceeded the limit for (2013). Additionally, the ANOVA also de- cereal, legumes and potato values. The total picts that there is no significant difference mean value (0.19 mg/kgfw) found in our between As content in crops (p>0.05, one- study is higher than the international regula- way). The t-test (0.05 level of significance) tions by ~ two fold. These Pb contents are finds that for As there is a significant differ- much lower than the data in the review da- ence between beans and potato. Additional- tabase (mean: 3.28 mg/kgfw; range: 0.03 – ly, TEs in potato skin and broad bean skin 43.5 mg/kgfw) reported by Rötting et al. have dietary significance when potato and (2013).

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

The statistical analyses were particularly in- groundwater can be related to the amount of teresting in one-way ANOVA (p< 0.05) for the As extracted in the steps 1 and 2, frac- As, Cd and Pb, which determined that there tions mobilized due to changes of pH and are no significant differences in mean values phosphate concentrations in the soils, which between crops for As an Pb, but there is a agree with the factor 5 that show a competi- significant difference for Cd contents be- tion between phosphate and arsenate for the tween crops. Additionally, Student´s t-test adsorption sites. All the surface stream sam- (at p = 0.05) for Cd found that there are ples located in the headwater of the basin are significant differences between mean con- dominated by baseflow and show a range centrations for beans versus barley and bar- pH of 9.1-12.3 (median11.6) (Papers II and ley versus potato crops. III). In contrast, along the area affected by AMD and ARD the pH range does not 7.6 Environmental index assessment show any significant difference between dry The results of the present work reveals that and rainy season (3.1 – 4.7; 3.3 – 5.2), re- the TEs can mobilize from soil-water (sur- spectively, in the Antequera sub-basin; in- face and groundwater) (Papers II, III & IV), stead the range is 3.7 – 7.2 in the Poopó soil-crop (edible part) (Paper IV) and sedi- sub-basin in the rainy season (Papers II & ment-water (Paper V). III). In surface water, comparison of Eh 7.6.1 The relationship between As and TEs values for both dry and rainy seasons in the in water (Papers I, II &III) Antequera and Poopó sub-basins shows that The groundwater show a nearly neutral pH the range does not seem to changes and the (7.2) around the PB and in the Antequera range is 117 – 381 mV (Papers I, II & III) in and Poopó sub-basins basic pH (median 8.1) areas affected AMD/ARD. The headwater (Paper I) and in both sub-basin moderately samples (before flowing through areas with reducing in some places, but oxidizing in mining activities) shows that the As concen- others (Papers II & III); this groundwater trations are < 20.5 µg/L, Cd are less than 2.5 characteristics govern the chemistry of TEs µg/L and in the thermal water As <45 µg/L around the studies areas. In PB, concentra- and Cd <1.2 µg/L. The surface water sam- tion of As was found in the range of 100 – ples in Antequera sub-basin indicate high 250 µg/L in the southern region (III) and in concentration of Zn>Fe>Cd>Mn>As>Cu. the northwestern region (V), whereas it was The high contamination of Cd could be ex- below 50 µg/L in the other regions (I, II and plained by the huge mining wastes aban- IV) and the concentrations of Cu, Cd, Pb doned along the Antequera River. The order and Zn are low in the regions around the of concentration for the Poopó sub-basin is PB, except in region II in the dry season Fe> Zn> As> Mn> Cd> Cu (Paper III). (Paper I). A detailed study in Antequera and The drinking water concentrations show Poopó sub-basins reveals that As, Cd and concentration for As < 17 µg/L, for Cd < Mn concentrations exceed World Health 1.1 µg/L and for Mn < 7.7 µg/L. The factor Organization (WHO) drinking guidelines analysis reveals that the mobilization of TEs (respective values: 10 µg/L, 3 µg/L and 400 is governed also by weathering of (plagio- µg/L; WHO, 2011), but Cu, Pb and Zn do clase, evaporates, calcites and gypsum) and not (Paper III). The factor analysis reveals oxidation of sulfides enhanced at the acidic that the mobilization of TEs in groundwater pH (~3.4) in Antequera and Poopó sub- is due to weathering of minerals (plagioclase, basins. The above presented indicates that in evaporates, calcites and gypsum) which re- studied areas in Bolivian Altiplano there is leased TEs, and desorption from hydrous an elevated risk to ingestion to TEs in both ferric oxide (HFO) surfaces (factor 4; Ante- drinking water and probably dermal expo- quera sub-basin) and competition between sure via recreational thermal waters. The 3- exposure to TEs most of them naturally As-PO4 (factor 5) in Antequera sub-basin - 4- present in the environments (e.g. As, Cd and and As-HCO3 and As-SiO4 (factor 3) in Poopó sub-basin. The As concentration in Mn) via ingestion (drinking water (Papers I,

Geochemistry of trace elements in the Bolivian Altiplano

Figure 26. Concept model of the occurrence of As and others TEs in groundwater, surface water and thermal springs in mining areas in Bolivia Altiplano.

II and III), food (Paper IV), inhalation (i.e. concluded that Cd is more chalcophile ele- dust) and dermal, can cause both non- ment than Zn and tends to persist as a sul- carcinogenic and carcinogenic health conse- fide while ZnS is weathering. Therefore, Cd quences to the human body (Kapaj et al became enriched relative to Zn in the soil 2006; Nordberg 2009; Hughes et al., 2011; which explains its relative immobility in sur- Hug et al., 2011; Halder, 2013 and ref. there face soils. in) (Fig. 26). Furthermore, it is worth considering that 7.6.2 Index with respect to TEs in soils and most of the As (%) in these soils was ex- crops (Paper IV) tracted with ammonium oxalate (fractions 3 For DTPA extracts, based on the definition – 4), possibly indicating the association with availability ratio (AR), the DTPA extraction Fe and Al oxides; meaning that these results show that less than 2% for As; <33% for Cd confirm the crucial role of Fe and Al oxides (except in one sample, 80%); <9% for Cu; in immobilizing As. However, high contri- <11% for Ni; <5% for Pb and <10% for butions of the fixed As fraction, (i.e. F3, F4, Zn (except in one sample, 48%) (Paper IV). F5) should be considered as beneficial from the aspect of environmental risk. The The soils show a low ratio of enrichment for amounts of As in fraction 2, which indicate Cu, Pb and Zn, meaning the availabilities of surface-bound As species and inner-sphere those TEs are limited; however, Cd concen- surface complexes of As constitute < 10% trations show a particular behavior with a of the total As, a potentially mobilized frac- high ratio from 11% to 33% indicating that tion, which could be transferred to crops or the plant-available fraction for Cd is relative- dissolved in groundwater. Therefore, the ly high. Tapia et al (2012) found that Cd environmental risk should be evaluated as concentrations are higher in Eastern Cordil- low. lera surficial soils, and Fuge et al. (1993)

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

The quality of irrigation water is also studied 7.6.4 Sequestration and mobility of TEs in in the present work, sodium adsorption ratio lake sediments (SAR) and percentage of sodium (Na%) Risk assessment code (RAC) (Paper V) determine the suitability of water for irriga- In general, the sediments show from low to tion purposes, and reveal that although moderate risk for Cu and Fe; moreover, many SAR´s values for surface water and moderate to high risk for Zn, Pb, Cd, Mn, groundwater are acceptable, high Na% of Co, and Ni, these values were determined the sampled surface waters is adverse for based on the percentage of the total metal crop growth. Even though some waters in content present in the first sediment fraction the Antequera sub-basin are composed of a (F1), and their presence may result in a seri- mixture of thermal water and runoff, Na% ous impact on the aquatic ecosystem with values are still too high for irrigation use. metals easily able to enter the food chain. In 7.6.3 The relationship between TEs in soil the S1 sample, Pb is not dissolved and in S2 and crops: assessment of transfer Cu, Zn, and Fe are strongly bound and not factor (Paper IV) easily mobilized into the aqueous phase. The transfer factor (TF) is an indicator of Sample S3 shows that Cu, Pb, and Fe pre- the amount of TEs in the edible part of sent a low risk for release into the water crop, calculated values show different ranges column. The S4 sampling site shows that Cu, for As (0.002–0.020), Cd (0.09–0.143), and Pb, Cd, and Ni present a low risk of dissolution in the water column, while heavy metals Pb (0.0003–0.02). The means values of TFAs are 2% for beans and barley, 0.8% for alfal- such as Zn, Fe and Mn present from moder- fa, and 0.6% for potato, indicating that only ate to high risk. The S5 shows low risk for a small fraction is transported from roots to Cu and Fe to be remobilized, but the risk for the edible part compared with the 10% Zn, Pb Cd, Fe, Mn and Ni is from moderate available As (in step 2, Paper IV) suggesting to high. The S6 sample shows a low risk for that As can be accumulated in the root ra- Cu, Pb, Cd, Fe and Ni and a higher risk for ther than in the edible parts. This is based Zn, Fe and Mn. The risk for Cu, Zn and Fe on finding that As can be accumulated in the in sample S7 is low, but there is a high risk soils due to amorphous and crystalline Fe for Pb, Cd, Mn, Co and Ni to enter into the oxides surfaces present in the soils, which human food chain through fish farming ac- was confirmed by the fixed arsenic deter- tivities near Tiquina town. The samples S8 mined fractions (fractions 3, 4, and 5) and and S9 show a similar behavior, the risk for also for the aerobic conditions favorable to Zn and Fe is low, while the other metals barley cultivation, where less mobile As(V) show from moderate to high risk to be mo- prevails with resulting retention of As bilized. Only Mn represents high and ex- (Williams et al., 2007). treme risk because it ca be remobilized by changes in environmental condition such as The mean values of TFCd are 6.8% for beans, pH and Eh. 2.2% for barley, 4.7% for potato and 3.8% for alfalfa. The uptake of Cd in presence of In general, the RAC values show that Cu high Cd available concentrations support the and Fe represent from low to moderately findings that agree with Oporto et al. (2009), risk; moreover, Cd, Co, Mn, Ni, Pb, and Zn indicate from moderate to high risk to hy- due to the relatively high available CdDTPA in the soils and low concentrations in barley drologic systems and thereafter entering into crops. the food chain (i.e. fish). Geoaccumulation index, Igeo (Paper V) The mean values of TFPb are much lower than those for Cd, varying from 0.9% for The Igeo was calculated to determine metal beans to 0.5% for barley, potato and alfalfa, contamination in sediments of Lake Titicaca indicating that much of the Pb available by comparing the observed concentrations could be fixed and retained in the roots (Pa- with reference values for crustal sediments per IV). (Fig. 27). Zinc has a particular distribution in

Geochemistry of trace elements in the Bolivian Altiplano the studied sediments because it is enriched and Pb, and finally extreme pollution for Cd. due to mining activities as show in samples Additionally, EF values suggest that Zn, Pb, S1, S4 and S5. All other sites are not consid- and Cd have extremely high enrichment in ered polluted. Lead varies from non-polluted the sediments in the point close to mining to highly polluted, again sample S4 has a activities (S4). Based on results of this study, higher value that reflects the impact of min- it is important to perform further studies to ing activities. For samples S1, S2, S3, S6 and determine possible bioaccumulation of S9 the Igeo values indicate a moderate pollu- heavy metals in the fish fauna since fish is a tion that can be attributed to exploitation of staple food of the local population. Pb, Zn, and Ag in polymetallic deposits. Cadmium in samples S1, S5 and S6 has val- 8 CONCLUSIONS ues that indicate from no pollution to mod- TEs study was conducted in three different erate pollution, sample S3 shows moderate spatial units: Basin, Sub-basin and transect in contamination, samples S2, S7, S8 and S9 mining areas. The importance of the Poopó show moderate to high pollution, and finally Basin approach was to assess the distribu- sample S4 is highly polluted which again tion of As and other TEs around the basin, could be attributed to the mining activities. but it was not possible identify the processes Metal enrichment factor, EF (Paper V) associated with each region. However, some processes were suggested on the basis of The EF values are higher than 1; the TEs field, lab investigation and modeling results. showing significant enrichment (5 – 20) are The high concentrations of As and TEs are Cu, Mn, Ni, and Zn, while Co varies be- not always associated with (mining) activities tween significant and very high enrichment in mining areas, these also can occur in areas in the shallow sediments. Conversely, Cd, Pb where there are no mining activities (south- and Zn are extremely enriched in the sites ern side of Lake Poopó), highlighting that with mining activities; whereas in other sites the characteristics of each site play an im- they are significantly enriched. Hence, the portant role in the mobilization and attenua- values higher than 1 indicated that all TEs tion of TEs, including both biotic and abiot- measured in the sediments were enrichment ic processes. by various anthropogenic sources in the basin area of Lake Titicaca. Hence, the RAC The present study shows that the hydro- evaluation shows from moderate to high risk chemistry of groundwater is controlled by for mobilization and/or remobilization of several processes; in the Antequera and Po- metals (Cu, Zn, Pb, Cd, Fe, Mn, Co and Ni) opó sub-basins there is a dominant silicate from the bottom sediment to the aquatic weathering; moreover, dissolution of car- ecosystem. There are both similarities and bonate minerals is more significant in the discrepancies between the Igeo and EF val- Antequera than in the Poopó sub-basin. ues, but both indicators show the enrich- Cation exchange is also partly responsible ment of Zn, Pb and Cd in sediments. How- for higher Na+ and K+ concentrations. ever, the discrepancies could be explained by However the halite dissolution is a more the variability of sediment sources, local important process. In surface water, the geology and the depth of the sediments and Gibbs diagram reveals that both rock weath- also by different assumptions implemented ering and evaporation-crystallization pro- in each environmental indicator calculation. cesses control water chemistry in the Ante- quera and Poopó samples. Moreover, the According to the Risk Assessment Code data suggest predominance of silicate weath- (RAC), there is a low to moderate risk for ering in both sub-basins followed by dissolu- Cu and Fe, while there is a moderate to high tion of evaporates and carbonate weathering, risk related to Zn, Pb, Cd, Mn, Co, and Ni additionaly the oxidation of sulfide minerals remobilization to the water column. The to some extent is also important in the study Igeo index indicates unpolluted status for area. Cu, Mn, Co, and Ni; high pollution for Zn

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

In both Antequera and Poopó sub-basins of FAO, WHO, EC, and Chilean interna- factor analysis was applied and it was possi- tional regulations. In contrast, Pb shows ble to identify five geochemical processes: i) high concentrations for beans and potato by plagioclase weathering, ii) dissolution of ~ two fold, but not for barley. gypsum and halite, iii) TE mobilization at The understanding of the TE in lake sedi- acidic pH, iv) release of As following comments and their mobilization to the water petition with phosphate and bicarbonate for column was improved by the sequential ex- adsorption sites. Whereas for surface water traction method. The BCR method was used data suggests the dominance of the follow- to evaluate the distribution of TEs in four ing processes: i) weathering and mobilization sediment fractions, the exchangeable, reduc- of TEs influenced by pH, ii) dissolution of ible, oxidizable and residual fractions. The evaporite salts, iii) neutralization of acid TE concentrations in the fractions agree mine drainage, iv) As release due to dissolu- with the information on chemical processes tion of Mn and Fe oxides, and v) sulfide that can explain the adsorption and mobili- oxidation. zation of TEs in lake sediments. A detailed transect study in soils was used for evaluating the pseudo total TE content 9 FUTURE WORK in soil, the data indicate that the geogenic In future work, it is important to understand signal is strong suggesting that the back- the processes that govern the dissolution of ground contents of these TEs are high. arsenic as a function of the water level However, in terms of bioavailable TEs con- through piezometer installation. Moreover centrations in soils follow the order Cd> sequential extractions for As in soils and Zn> Cu> Pb, and the distribution of bioa- sediments may be modified to consider the vailable Cd and AR values suggest continu- salinity of the groundwater (due to i.e. the ous soil enrichment. The DTPA method dissolution of halite). extracted show less than 2% of the total As, while the sequential methods reported up to It is also important to consider the specia- 12% (< 2%, step 1 and < 10.0%, step 2), tion of TEs (especially organic-As and inor- which represents less than 3.1 mg/kg of the ganic-As) in crops and their bioaccessibility total As content, as a potentially mobilized for the human body, linked to WHO/FAO fraction, which could be transferred to crops provisional tolerable weekly intake (PTWI). or dissolved in groundwater. The large pool TEs should be further explored in mining of As can be accumulated in the soils due to areas of the Bolivian Altiplano for minimiz- amorphous and crystalline Fe oxide surfaces ing the potential health risks. present in the soils, which was confirmed by Future research based on the WHO/FAO the fixed arsenic fraction (fractions 3, 4, and provisional tolerable weekly intake (PTWI) 5). of trace metals is needed in the study areas Furthermore, evaluation of the edible part of of the Bolivian Altiplano affected by mining the crops showed that As and Cd concentra- to minimize the potential health risks from tions in the edible part of the crops (beans, TE contamination in the food chain. barley and potato) are lower than the limits

Geochemistry of trace elements in the Bolivian Altiplano

10 REFERENCES Allison JDD, Brown S, Novo-Gradac KJ. 1991. MINTEQA2, a geochemical assessment data base and test cases for environmental systems. EPA, Athens, GA, U.S. Anawar HM, Akai J, Sakugawa H, 2004. Mobilization of arsenic from subsurface sediments by effect of bicarbonate ions in groundwater. Chemosphere. 54:753-762. Appelo CAJ, Postma D. 2005. Geochemistry, groundwater and pollution. Balkema, Rotterdam, NL, 536 p. Aravena R. 1995. Isotope hydrology and geochemistry of northern Chile groundwaters. Bull. Inst. Fr. Etudes Andines. 24(3):495-503. Archer J, Hudson-Edwards KA, Preston DA, Howarth R, Linge K. 2005. Aqueous exposure and uptake of arsenic by riverside communities affected by mining contamination in the Río Pil- comayo basin, Bolivia. Mineralogical Magazine. 69(5):719-736. Arce-Burgoa O, Goldfarb R. 2009.Metallogeny of Bolivia. Society of Economic Geologist Newletter, 79p. Argollo J, Mourguiart P. 2000. Late quaternary climate history of the Bolivian Altiplano. Quater- nary International. 72:37-51. Aprile FM, Bouvy M. 2008. Distribution and enrichment of heavy metals in sediments in the Tapacurá River Basin. Northeastern Brazil. Brazilian Journal of Aquatic Sciences and Technology. 12(1):1-8. ASA (American Society of Agronomy). 1982. Methods of Soil Analysis. Part 2. American Society of Agronomy, Madison, WI. Bacon J, Davidson C. 2008. Is there a future for sequential chemical extraction? Analyst. 133:25- 46. Baker PA, Seltzer GO, Fritz SC, Dunbar RB, Grove MJ, Tapia PM, Cross SL, Rowe HD, Broda JP, 2001. The history of tropical South American precipitation for the past 25,000 years. Sci- ence. 291:640-643. Banks D, Markland H, Smith PV, Mendez C, Rodriguez J, Huerta A, Saether OM. 2004 Distribu- tion, salinity and pH dependence of elements in surface water of the catchment area of the Salars of Coipasa and Uyuni Bolivian Altiplano. J. Geochemistry Exploration. 84:141-166. Banks D, Holden W, Aguilar E, Mendez C, Koller D, Andia Z, Rodriguez J, Saether O, Torrico A, Veneros R, Flores J. 2002. Contaminant source characterization of the San José Mine, Oru- ro, Bolivia. In: Mine water hydrogeology and geochemistry, Younger PL, Robin NS (eds). Geological Society, London, special publications. 198:215-239. Banks D, Markland H, Smith PV, Mendez C, Rodriguez J, Huerta A, Saether OM. 2004. Distri- bution, salinity and pH dependence of elements in surface water of the catchment area of the Salars of Coipasa and Uyuni Bolivian Altiplano. Journal Geochemistry Exploration. 84:141-166. Batista BL, de Oliveira Souza VC, da Silva FG, Barbosa F. 2010. Survey of 13 trace elements of toxic and nutritional significance in rice from Brazil and exposure assessment. Food Additives and Contaminants: Part B. 3:253-262. Bermudez GMA, Jasan R. Plá R, Pignata M. 2012. Heavy metals and trace elements in atmospheric fall-out: Their relationship with topsoil and wheat element composition. Journal of Haz- ardous Materials. 213-214:447-456. Bergqvist C, Herbert R, Persson I, Greger M. 2014. Plants influence on arsenic availability and speciation in the rhizosphere, roots and shoots of three different vegetables. Environmental Pol- lution. 184:540-546.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Bhattacharya A, Routh J, Jacks G, Bhattacharya P, Mörth M. 2006a. Environmental assessment of abandoned mine tailing in Adak, Väterbotten district (northern Sweden). App. Geochem. 21:1760-1780. Bhattacharya P, Claesson M, Bundschuh J, Sracek O, Fagerberg J, Jacks G, Martin RA, Storniolo A,Thir J. 2006b. Distribution and mobility of arsenic in the Rio Dulce Alluvial aquifer in San- tiago del Estero Province, Argentina. Sci. Total Environ. 358:97-120. Bothe JV, Brown PW. 1999. The stabilities of calcium arsenates. Journal of Hazardous Materials. 69(2):197-207. Biswas A, Nath B, Bhattacharya P, Halder D, Kundu AK, Mandal U, Mukhrjee A, Chatterjee D, Mörth C-M, Jacks G. 2012. Hydrogeochemical constrast between brown and grey aquifer in shallow depth of Bengal Basin: Consequences for sustainable drinking water supply. Sci. Total Environ. 431:402-412. Boulange B, Aquize JE, 1981. Morphologie, hydrographie et climatologie du lac Titicaca et de son basin versant. Rev. Hydrobiol.Trop. 14(4):269-287. Bundschuh J, Farias B, Martin R, Storniolo A, Bhattacharya P, Cortes J, Bonorino G, Albouy R. 2004. Groundwater arsenic in the Chaco-Pampean plain, Argentina: case study from Robles County, Santiago del Estero Province. Appl. Geochem. 19:231-243. Bundschuh J, Litter MI, Parvez F, Román-Ross G, Nicolli HB, Jean J, Liu C, Lopez D, Armienta MA, Guilherme LRG, Gomez 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. Sci. Total Environ. 429:2-35. Burak DL, Fontes M, Terra Santos N, Soares LV, De Sousa E, Becquer T. 2010. Geochemistry and spatial distribution of heavy metals in oxisols in a mineralized region of the Brazilian Cen- tral Plateau. Geoderma. 160:131-142. Cáceres Choque LF, Choque RR, Ramos Ramos OE, Choque R. 2004. Trace elements speciation in sediments of Lake Poopó (Especiación de elementos traza en sedimentos del Lago Poopó). Revista Boliviana de Química. 21(1):42-48. Cáceres Choque LF, Ramos Ramos OE, Valdez SN, Choque RR, Choque RG, Fernández SG, Sracek O, Bhattacharya P. 2013. Fractionation of heavy metals and assessment of contamination of the sediments of Lake Titicaca. Environmental Monitoring Assessment Journal. 185:9979- 9994. Carmouze JP, Arze C, Quintanilla J. 1981. Régulation hydroclimique du lac Titicaca et l'hydro- chimie de ses tributaires. Rev. Hydrobiol. Trop. 14:329-348. Carmouze J, Arze C, Quintanilla J. 1992. Hydrochemical regulation of the lake and water chemistry of its inflow rivers. In: Lake Titicaca: A synthesis of limnological knowledge, Dejoux C, Iltis A. (eds) Kluwer Academic Publishers. 98-112. Calizaya A. 2009. Water resource management efforts for best water allocation in the lake Poopó Basin, Bolivia. Lund University. Lund, Sweden. Doctoral thesis. Coudrian-Ribstein A, Pratx B, Quintanilla J, Zuppi GM, Cahuaya D. 1995a. Salinity of the groundwater on the Central Altiplano. In: Ribstein P, Francou B, Coudrain-Ribstein A, Mour- guiart P. (eds.). Bulletin de l’ Institut Français d’Études Andines. 24(3):483-493. Coudrian-Ribstein A, Olive P, Quintanilla J, Sondag F, Cahuaya D. 1995b. Salinity and isotopic dynamics of the groundwater resources on the Bolivian Altiplano. Application of tracers in arid zone hydrology (Proceedings of Vienna Symposium, 1994). IAHS Publ. 232. Cui Y-J, Zhu Y-G, Zhai R-H, Chen D-Y, Huang Y-Z, Qiu Y, Liang J-Z. 2004. Transfer of metals from soil to vegetables in an area near a smelter in Nanning, China. Environ Int. 30(6):785-791. Deschamps E, Matschullat J. 2007. Arsênio antropogênico e natural: um estudo em regiões do Quadilátero Ferrífero. Belo Horizonte 330 p.

Geochemistry of trace elements in the Bolivian Altiplano

Dejoux C, Iltis A. 1992. Lake Titicaca: A synthesis of limnological knowledge. Kluwer Academic Publis- hers. Díaz O, Leyton I, Muñoz O, Nuñez N, Devesa V, Súñer M, Vélez D, Montoro R. 2004. Contri- bution of water, bread and vegetables (raw and cooked) to dietary intake of inorganic arsenic in a rural village of northern Chile. Journal of Agricultural and Food Chemistry. 52:1773-1779. Díaz V. 2013a. De la fundación de la república al triunfo de la oligarquía minera; libre comercio de plata y abolición de la moneda feble. CEDIB. 29-33. Díaz V. 2013b. Siglo XIX auge y decadencia de la minería de la plata. Periodo 1872-1900. CE- DIB. 29-33. Dieter HH, Bayer TA, Multhaup G. 2005. Environmental copper and manganese in the patho- physiology of neurologic diseases (Alzheimer´s disease and Manganism). Acta Hydroch-Hydrob. 33:72-78. Dzombak D, Morel FM, 1990. Surface complexation modeling, hydrous ferric oxide. Wiley, New York. Commission Regulation (EC). 2006. Setting maximum levels for certain contaminants in food- stuffs #1881. Ettler V, Mihaljevic M, Sebek O, Molek M, Grygar T, Zeman J. 2006. Geochemical and Pb isotopic evidence for sources and dispersal of metal contaminations in stream sediments from the mining and smelting district of Príbam, Czech Republic. Environmental Pollution. 142:409- 417. Farnfield HR, Marcilla AI, Ward NI. 2012. Arsenic speciation and trace element analysis of the volcanic río Agrio and the geothermal waters of Copahue, Argentina. Sci. Total Environ. 433:371-378. Fontes JC, Boulange B, Carmouze JP, Florkowski T. 1979. Preliminary oxygen-18 and deuterium study of the dynamics of Lake Titicaca. Isotopes in Lake Studies. IAEA publ.145-150. Fornari M, Risacher F, Féraud G. 2001. Dating of paleolakes in the central Altiplano of Bolivia. Paleogeogr Paleoclimatol Paleoecol. 172:269-282. Förstner U. 1995. Land contamination by metals: global scope and magnitude of problem. In: Metal speciation and contamination of soil, Allen H, Huang C, Bailey G, Bowers A (eds) Lewis Pub- lishers. 1-24. Fuge R, Pearce FM, Pearce NJG, Perkins W. 1993. Geochemistry of Cd in the secondary environment near abandoned metalliferous mines, Wales. Appl. Geochem. 2:29-35. Garreaud R, Vuille M, Clement A. 2003. The climate of the Altiplano: observed current condition and mechanisms of past changes. Palaeogeography, Palaeoclimatology, Palaeoecology. 194:5-22. Garrett RG. 2004. Natural distribution and abundance of elements. In: Essentials of medical geology. Impacts of the natural environment on public health, Selinus O, Alloway B, Centeno J, Finkelman R, Fuge R, Lindh U, Smedley P (eds). Elsevier Academic Press. 17-41. GEOBOL, 1992. Carta geológica, Hoja Poopó. SGB Serie I-CGB-13. Gibbs RJ. 1970. Mechanism controlling world water chemistry. Science. 170(3962):1088-1090. Goix S, Point D, Oliva P, Polve M, Duprey JL, Mazurek H, Guislain L, Huayta C, Barbieri F, Gardon J. 2011. Influence of source distribution and geochemical composition of aerosols on children exposure in the large polymetallic mining region of the Bolivian Altiplano. Sci. Total Environ . 412 – 413:170-184. González E, Retamal M, Medina V, Ahumada R, Neira J. 2009. Enriquecimiento, disponibilidad y contaminación de metals traza (Cd, Cu, Pb y Zn) en sedimentos de lagunas urbanas de Con- cepción-Chile. Química Nova. 32(4):902-907.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Grove M, Baker P, Cross S, Rigsby C, Seltzer G. 2003. Application of strontium isotopes to understanding the hydrology and paleohydrology of the Altiplano, Bolivia – Peru). Palaeogeogra- phy, Palaeoclimatology, Palaeoecology. 194:281-297. Halder D. 2013. Arsenic exposure risk from rice and other dietry components in rural Bengal. Doctoral Thesis. Land and Water Resources Engineering, Stockholm, Sweden. 33p. Hudson-Edwards KA, Macklin MG, Miller JR, Lechler PL. 2001. Sources, distribution and stor- age of heavy metals in the Río Pilcomayo, Bolivia. Journal of Geochemical Explorations. 72:229- 250. Hudson-Edwards KA, Houghton SL, Osborn A. 2004. Extraction and analysis of arsenic in soils and sediments. Trends in Analysis Chemistry. 23(10-11):745-752. Hug SJ, Gaertner D, Roberts L, Schirner M, Ruettiman T, Rosenberg TM, Badruzzaman ABM, Ali MA. 2011. Appl. Geochem. 26:1077-1085. Hughes M, Beck BD, Chen Y, Lewis AS, Thomas DJ. 2011. Arsenic exposure and toxicology: A historical perspective. Toxicological Sciences. 123(2):305-332. Jackson M. 1958. Soil Chemical Analysis. 1st ed. Prentice Hall Pub. Co., New York. Janhowski J, Acworth RI, Shekarforoush S. 1998. Reverse ion exchange in deeply weathered por- phyritic dacite fractured aquifer system. Yass, New South Wales, Australia. In: 9th Internat symp water-rock interaction, Arehart GB, Hulston JR, (eds) Rotterdam: A. A. Balkema; 213-246. Jacks G, Sefe F, Carling M, Hammar M, Letsamao P. 1999. Tentative nitrogen budget for pit latrine. Environ. Geol. 38:199-203. JECFA. 2005. Joint FAO/WHO expert committee and food additives. Sixty-fourth meeting. www.fao.org/es/esn/jecfa/index_en.stm. Accessed at November 28, 2012. Iltis A, Carmouze JP, Lemoalle J. 1992. Physico-chemical properties of the water. In. Lake Titica- ca: A synthesis of limnological knowledge, Dejoux C, Iltis A. (eds) Kluwer Academic Publishers. 89- 97. Kabata- Pendias A, Mukherjee AB. 2007. Trace elements from soil to human. Springer -Verlag Berlin Heidelberg New York Kabata-Pendias A, Pendias H. 2001.Trace elements in soils and plants. 3rd ed. CRC press LLC. Kapaj S, Peterson H, Liber K, Bhattacharya P. 2006. Human health effects from chronic arsenic poisoning: a review. J. Environ. Sci. Health A Tox. Hazard Subts. Environ. Eng. 41:2399-2428. Karageorgis AP, Katsanevakis S, Kaberi H. 2009. Use of enrichment factor for the assessment of heavy metal contamination in the sediments of Koumoundourou lake. Greece. Water Air Soil Pollut. 204:243-258. Khan K, Lu Y, Khan H, Ishtiaq M, Khan S, Waqas M, Wei L, Wang T. 2013. Heavy metals in agricultural soils and crops and their health risks in Swat District, northern Pakistan. Food and chemical toxicology. 58:449-458. Kossoff D, Hudson-Edwards KA, Dubbin WE, Alfredsson MA. 2011. Incongruent weathering of Cd and Zn from mine tailing: A column leaching study. Chemical Geology. 281:52-71. Kossoff D, Hudson-Edwards KA, Dubbin WE, Alfredsson M. 2012a. Major and trace metal mobility during weathering of mine tailing: Implication for floodplains soils. Appl. Geochem. 27:562-576. Kossoff D, Hudson-Edwards KA, Dubbin WE, Alfredsson M, Geraki T. 2012b. Cycling of As, P, Pb and Sb during weathering of mine tailing: implications for fluvial environments. Miner- alogical Magazine. 76(5):1209-1228. Lavenu A. 1992. Formation and geological evolution. In: Lake Titicaca: A synthesis of limnological knowledge, Dejoux C, Iltis A (eds) Kluwer Academic Publishers. 3-15.

Geochemistry of trace elements in the Bolivian Altiplano

Law 1333, 1992. Bolivian Environmental Law. Reglamento de la Ley 1333 del Medio Ambiente, Reglamento en materia de contaminación hídrica. Lilja A, Linde G. 2006. Occurrence and distribution of heavy metals in three rivers on the Bolivi- an high plateau. Lund Institute of Technology, Lund University, Lund. MSc thesis. Litter M, Pérez – Carrera A, Morgada ME, Ramos O, Quintanilla J, Fernandez – Cirelli A. 2008. Formas presentes de arsénico en agua y suelo. In: Distribución del arsénico en las regiones ibéricas e iberoamericanas, Bundschuh J, Pérez–Carrera A, Litter M (eds), Cyted-Argentina. 5-32. Lizarazu J, Aranyossy J, Orsag V, Salazar JC. 1987. Estudio isotópico de la cuenca de Oruro- Caracollo (Bolivia). Isotope Techniques in water resources development (Pro. Vienna symposium March 1987), IAEA publ. 301-314. McArthur JM, Ravenscroft P, Banerjee DM. Milsom J, Hudson-Edwards KA, Sengupta S, Bris- tow C, Sarkar A, Tonkin S, Purohit R., 2008. How paleosols influence groundwater flow and arsenic pollution: A model from the Bengal Basin and its worldwide implication. Water Resour. Res. 44, W11411. Massas I, Ehaliotis C, Geronditis S, Sarris E. 2009. Elevated heavy metal concentrationes in tops soils of an Aegean island town (Greece): total and available forms, origin and distribution. En- viron Monit Assess. 151:105-116. Massas I, Ehaliotis C, Kalivas D, Panagopoulou G. 2010. Concentrations and availability indicator of heavy metals; the case of children´s playground in the city of Athens (Greece). Water, Air and Soil Pollution. 212(1-4):51-63. Massas I, Kalivas D, Ehaliotis C, Gasparatos D. 2013. Total and available heavy metal concentrations in soil of the Thriassio plan (Greece) and assessment of soil pollution indexes. Environ Monit Assess. 185:6751-6766. Mandal B, Suzuki K. 2002. Arsenic round the world: a review. Talanta. 58:201-235. Matschullat J. 2000. Arsenic in the geosphere–a review. The Science of the Total Environment. 249:297-312. McBride M.B. 1994. Trace and toxic elements in soils. In: Environmental chemistry of soils. Oxford University Press, New York. 308-341. Medeiros RJ, dos Santos LM, Freire AS, Santelli RE, Braga AM, Krauss TM, Jacob S. 2012. De- termination of inorganic trace elements in edible marine fish from Rio de Janeiro State, Brazil. Food Control. 23:535-541. Miller JR, Hudson-Edwards KA, Lechler PJ, Preston D, Macklin MG. 2004. Heavy metal contamination of water, soil and produce within riverine communities of the Río Pilcomayo basin, Bolivia. Science of the Total Environment. 320:189-209. Ministerio de Minería y Metalurgia (MMM) 2012. Propuesta plan sectorial 2012. Ministerio de Minería y Metalurgia (MMM) 2013. Anuario estadístico minero del Ministerio de Minería y Metalurgía del 2013. Minchin J. 1882. Notes on a journey through part of the Andean tableland of Bolivia. Proc R. Geogr. Soc. 4:67. Molina M, Aburto F, Calderon R, Cazanga M, Escudey M. 2009. Trace element composition of selected fertilizers used in Chile: Phosphorus fertilizers as a source of long-term soil contamination. Soil and Sediment Contamination. 18:497-511. Morel MMF. 1983. Principles of aquatic chemistry. John Wiley and Sons Publishers. Somerset, N.J. 446p. Morgano MA, Rabonato LC, Milani RF, Miyagusku L, Balian S. 2011. Assessment of trace elements in fishes of Japanese foods marketed in Sao Paulo (Brazil). Food Control. 22:778-785.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Mourguiart P, Wirrmann D, Fournier M, Servant M. 1992. Reconstruction quantitative des niveaux du petit lac Titicaca au cours de l´Holocene. C.R. Acad. Sci., Paris. 315:875-880. Mourguiart P. 2000. Historical changes in the environment of Lake Titicaca: evidence from os- tracod ecology and evolution. Advances in ecological research. 31:497-520. Muñoz O, Bastias JM, Araya M, Morales A, Orellana C, Rebolledo R, Velez D. 2005. Estimation of the dietary intake of cadmium, lead, mercury, and arsenic by the population of Santiago (Chile) using a total diet study. Food and Chemical Toxicology. 43:1647-1655. Muñoz O, Díaz O. Leyton I, Nuñez N, Devesa V, Súñer M, Vélez D, Montoro R. 2002. Vegeta- bles collected in the cultivated Andean area of northern Chile: Total and inorganic arsenic contents in raw vegetables. Journal of agricultural and food chemistry. 50:642-647. Muhammad S, Shah MT, Khan S, 2011. Health risk assessment of heavy metals and their source apportionment in drinking water of Kohistan region, northern Pakistan. MicroChem. J. 98:334- 343. Mukherjee A, von Brömssen M, Scanlon BR, Bhattacharya P, Fryar A, Hasan MA, Ahmed KM, Chatterjee D, Jacks G, Sracek O. 2008. Hydrogeochemical comparison and effects of overlap- ping redox zones on groundwater arsenic near the Western (Bhagirathi sub-basin, India) and Eastern (Meghna sub-basin, Bangladesh) margins of the Bengal basin. Journal of Contaminant Hydrology. 99:31-48. Mukherjee A, Bhattacharya P, Shi F, Fryar A, Mukherjee AB, Xie ZM, et al. 2009. Chemical evolution in the high arsenic groundwater of the Huhhot basin (Inner Mongolia, PR China) and its difference from the western Bengal basin (India). Appl. Geochem. 24:1835-1851. Müller G. 1981. Die Schwermetallbelastung der sedimenten des Neckars und seiner Nebenflüsse: eine Bestandsaufnahme. Chemiker-Zeitung. 105:157-164. Niazi NK, Singh B, Shah P. 2011. Arsenic speciation and phytoavailability in contaminated soils using a sequential extraction procedure and XANES spectroscopy. Environmental Science and Technology. 45:7135-7142. Neaman A, Huerta S, Sauvé S. 2012. Effects of lime and compost on earthworm (Eisenia fetida) reproduction in copper and arsenic contaminated soils from the Puchuncaví Valley, Chile. Eco- toxicology and Environmental Safety. 80:386-392. Nicolli HB, Bundschuh J, Blanco M, Tujchneider O, Panarello HO, Dapeña C, Rusansky JE. 2012. Arsenic and associated trace-elements in groundwater from the Chaco-Pampean plan, Argentina: Results from 100 years of research. Science of the Total Environment. 429:36-56. Nicolli HB, Bundschuh J, García JW, Falcón CM, Jean J. 2010. Sources and controls for the mobility of arsenic in oxidizing groundwater from loess-type sediments in arid/semi-arid dry cli- mates – Evidence from the Chaco-Pampean plain (Argentina). Water Research. 44:5589- 5604. Nicolli HB, García JW, Falcón CM, Smedley PL. 2012b. Mobilization of arsenic and other trace elements of health concern in groundwater from the Salí River basin Tucumán Province, Ar- gentina. Environmental Geochemistry and Health. 34:251-262. Nittetsu Mining Consultants CO., LTD, and Leman Consulting Group, 2012. Estudio sobre la optimizacion en el uso del agua subterránea en la mina San Cristóbal y el agua subterránea en el área circundante. Nordberg G, Jin T, Bernard A, Fierens S, Buchet JP,Ye T, Kong Q, Wang H. 2002. Low bone density and renal dysfuntion following environmental cadmium exposure in China. Ambio. 31:478-481. Nordberg G. 2009. Historical perspective on cadmium toxicology. Toxicology and applied pharmacol- ogy. 238:192-200. NB-512, 2004. Reglamento Nacional para el Control de la Calidad del Agua para Consumo Hu- mano. Ministerio de Servicios y Obras Públicas, Vice-Ministerio de Servicios Básicos, 59 p.

Geochemistry of trace elements in the Bolivian Altiplano

O`Reilly J, Watts MJ, Shaw RA, Marcilla AL, Ward NI. 2010. Arsenic contamination of natural waters in San Juan and La Pampa, Argentina. Environmental Geochemistry and Health. 32:491-515. Ono FB, Guimaraes LR, Silva E, Santos G, Hale B, Toujaguez R, Bundschuh J. 2012. Arsenic bioaccessibility in a gold mining area: a health risk assessment for children. Environmental Geo- chemistry and Health. 34:457-465. Oporto C, Vandecasteele C, Smolders E. 2007. Elevated cadmium concnetrations in potato tu- bers due to irrigation with river wáter contamined by mining in Potosí, Bolivia. J. Environ. Qual. 36:1181-1186. Oporto C, Smolders E, Degryse F, Verheyen L, Vandescasteele C. 2009. DGT-measured fluxes explain the chloride enhanced cadmium uptake by plants at low but not a high Cd supply. Plant Soil. 318:127-135. PAADO. 2005. Plan de Adecuación Ambiental del Departamento de Oruro. Prefectura de Oru- ro, Bolivia. Parkhurst DL, Appelo CAJ. 1999. Users´s guide to PHREEQC (version 2)-A computer program for speciation, batch reaction, one-dimensional transport and inverse geochemical calculations. U.S. Geological Survey, Denver, Colorado, USA. 99 – 4259. Passos E, Alves J, dos Santos I, Alves JP, Garcia C. Spinola Costa A. 2010. Assessment of trace metals contamination in estuarine sediments using a sequential technique and principal component analysis. Microchemical Journal. 96:50-57. Pempkowiase J, Sikora A, Biernacka E. 1999. Speciation of heavy metals in marine sediments versus their accumulation by mussels. Chemosphere. 39(2):313-321. Pérez-Carrera A, Fernández-Cirelli A. 2005. Arsenic concentration in water and bovine milk in Cordoba, Argentina. Preliminary results. Journal of Dairy Research. 72:122-124. Petersen W, Willer E, Willamowski C. 1997. Remobilization of trace elements from polluted anoxic sediments after resuspension in oxic water. Water, Air, and Soil Pollution. 99:515-522. Pillco R, Bengtsson L. 2006. Long-term and extreme water level variation of the shallow Lake Poopó, Bolivia. Hydrological Sciences Journal. 51:98-114. PPO–03, Plan Piloto Oruro – 03, 1996. Panorama de la fisiografía y geología del área del proyec- to. Ministerio de Desarrollo Sostenible y Medio Ambiente, Swedish Geological AB, La Paz, Bolivia. PPO–04, Plan Piloto Oruro – 04, 1996. Evaluación de recursos minerales y su utilización. Minis- terio de Desarrollo Sostenible y Medio Ambiente, Swedish Geological, La Paz, Bolivia. PPO–13, Plan Piloto Oruro – 013. 1996. Metal contents in lake sediments, totora and myriophyl- lum of Lake Uru Uru. Ministerio de Desarrollo Sostenible y Medio Ambiente, Swedish Geo- logical AB, La Paz, Bolivia, 128 p. Prieto-García F, Callejas HJ, Lechuga M de los Á, Gaytán JC, Barrado EE. 2005. Accumulation in vegetable weavings of arsenic originating from water and floors of Zimapán, Hidalgo State, Mexico. Bioagro. 17(3):129-136. Parkhurst DL, Appelo CAJ. 1999. User’s guide to PHREEQC (Version 2) - a computer program for speciation, batch-reaction, one-dimensional transport, and inverse geochemical calculations, Report of investigation - U.S. Geological Survey and Water Resources. 312p. Queirolo F, Stegen S, Restovic M, Paz M, Ostapczuk P, Schwuger MJ, Muñoz O. 2000. Total arsenic, lead, cadmium level in vegetables cultivated at the Andean villages of northern Chile. The Science of the Total Environment. 255:75-84. Quino I. 2006. Groundwater quality determination on the northern and eastern regions of the Lake Poopó (Determinación de la calidad de aguas subterráneas en la región norte y este del lago Poopó). Thesis, UMSA, La Paz, Bolivia.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Raab A, Ferreira K, Meharg AA, Feldman J. 2007. Can arsenic-phytochelatin complex formation be used as an indicador for toxicity in Helianthus annuus?. J. Exp. Bot. 58:1333-1338. Ravenscroft P, McArthur JM. 2004. Mechanism of regional enrichment of groundwater by bo- ron, USA. Appl. Geochem. 19:1412-1430. Revenga JE, Campbell LM, Arribére MA, Ribeiro S. 2012. Arsenic, cobalt and chromium food web biodilution in a Patagonia mountain lake. Ecotoxicology and Environmental Safety. 81:1-10. Redwood S. 1993. The metallogeny of the Bolivian Andes: Vancouver, British Columbia. Univer- sity of British Columbia. Mineral Deposit Research Unit. 15, 55p. Rigsby C, Bradbury P, Baker P, Rollins S, Warren M. 2005. Late quaternary palaeolakes, rivers and wetlands on the Bolivian Altiplano and their palaeoclimatic implications. Journal of Quater- nary Science. 20(7-8):671-691. Risacher F, Fritz B. 1991. Geochemistry of Bolivian salars, Lipez, southern Altiplano: Origin of solutes and brine evolution. Geochimica et Coscochimica Acta. 55:687-705. Roche M, Bourges J, Cortes J, Mattos R. 1992. Climatology and Hydrology of the Titicaca basin. In: Lake Titicaca: A synthesis of limnological knowledge, Dejoux C, Iltis A. (eds) Kluwer Academic Publishers. 63-88. Rodrigo L, Wirrmann D. 1992. General aspects of present-day sedimentation. In: Lake Titicaca: A synthesis of limnological knowledge, Dejoux C, Iltis A. (eds) Kluwer Academic Publishers. 23-28. Rosso JJ, Troncoso JJ, Fernández Cirelli A. 2011. Geographic distribution of arsenic and trace metal in lotic ecosystem of the pampa plain, Argentina. Bulletin of Environmental Contamination and Toxicology. 86:129-132. Rötting TS, Cama J, Ayora C, Cortina J-L, De Pablo J. 2006. Use of caustic Magnesia to remove cadmium, nickel, and cobalt from water in passive treatment systems: column experiments. Environmental Science & Technology. 40(20):6438-6443. Rötting TS, Mercado M, García ME, Quintanilla J. 2013. Environmental distribution and health impacts of As and Pb in crops and soils near Vinto smenter, Oruro, Bolivia. Int. J. Environ. Sci. Technol. DOI 10.1007/s13762-013-0313-1 Scarpelli W. 2005. Arsenic in the rivers of the Amazon basin. Terrae. 2(1-2):20-27. Selander L, Svan P. 2007. Occurrence and distribution of heavy metals in lake Poopó, Bolivia. Department of Chemical and Water Resources Engineering. Lund University, Lund. Master thesis SERGEOTECMIN, 1999. Inventariación de recursos naturales renovables (hídricos) y no renovables (minerales e hidrocarburos) del departamento de Oruro. Boletín del Servicio Nacional de Geología y Minería. 24, 44 p. Servant M, Fontes JC. 1978. Les lacs quaternaries des hauts plateaux des Andes boliviennes. Mountain Res. Develop. 13:129-138. Schiff KC, Weisberg, 1999. Iron as a reference element for determining trace elements enrich- ments in southern California coastal shelf sediments. Marine Environmental Research. 48:161-176. Signes-Pastor A, Burló F, Mitra K, Carbonell-Barrachina AA. 2007. Arsenic biogeochemistry as affected by phosphorus fertilizer addition, redox potential and pH in a west Bengal (India) soil. Geoderma. 137(3-4):504-510. Shaw AJ. 1990. Heavy metal tolerance in plants: Evolutionary aspects CRC PRes, Boca Raton, FL. Smedley PL, Kinniburgh DG. 2002. A review of the source, behavior and distribution of arsenic in natural waters. Appl. Geochem. 17:517-568.

Geochemistry of trace elements in the Bolivian Altiplano

Smedley PL, Kinniburgh DG, Macdonald DMJ, Nicolli HB, Barros AJ, Tullio JO, Pearce JM, Alonso MS. 2005. Arsenic associations in sediments from the loess aquifer of La Pampa, Ar- gentina. Appl. Geochem. 20:989-1016. Smolders AJP, Hudson-Edwards KA, Van der Velde G, Roelofs JGM. 2004. Control on water chemistry of the Pilcomayo river (Bolivia, South-America). Appl. Geochem. 19:1745-1758. Sposito G. 1989. The chemistry of soils. Oxford University Press, New York, 267 p. Sracek O, Bhattacharya P, Jacks G, Gustafsson JP, von Brömssen M. 2004. Behavior of arsenic and geochemical modeling of arsenic enrichment in aqueous enrironments. Appl. Geochem. 19:169-180. Stumm W, Morgan TT. 1996. Aquatic chemistry: Chemical equilibria and rate in natural waters. Third Edition Wiley Interscience. John Wiley and Sons, Inc; New York. 1022 p. Tapia P, Fritz S, Baker P, Seltzer G, Dunbar R. 2003. A late Quaternary diatom record of tropical climatic history from Lake Titicaca (Peru and Bolivia). Palaeogeography, Palaeoclimatology, Palaeoe- cology. 194:139-164. Sun J, Changyuan T, Wu P, Strosnoder WHJ, Han Z. 2013. Hydrogeochemical characteristics of streams with and without acid mine drainage impacts: A paired catchment study in karst geology, SW China. Journal of Hydrology. 504:115-124. Tapia J. 2011. Sources, mobility and bioavailability of metals and metalloids in the mining and smelter impacted altiplanean city of Oruro, Bolivia. Université de Toulouse. Doctoral thesis. 263p. Tapia J, Audry S, Townley B, Duprey JL. 2012. Geochemical background, baseline and origin of contaminants from sediments in the mining-impacted Altiplano and Eastern Cordillera of Oruro, Bolivia. Geochem. Explor. Environ. Anal. 12:3-20. Tapia J, Audry S. 2013. Control of early diagenesis processes on trace metal (Cu, Zn, Cd, Pb and U) and metalloid (As, Sb) behaviors in mining- and smeting-impacted lacustrine environments of the Bolivian Altiplano. Appl. Geochem. 31:60-78. TDPS. 1993. Climatología de los Sistemas Lago Titicaca, Lago Poopó, Desaguadero, Salares Coipasa y Uyuni (TDPS). Climatology of joint system Titicaca and Poopó Lakes, De- saguadero, Coipasa and Uyuni salts pan). Comisión de Comunidades Europeas – Republicas de Peru y Bolivia, convenios ALA/86/03, ALA/86/23, LP. Bolivia. Turdogan MK, Fevzi K, Kazim K, Ilyas T, Isamil U. 2003. Heavy metals in soils, vegetables and fruits in the endemic upper gastrointestinal cancer region of Turkey. Environ. Toxicol. Pharmacol. 13:175-179. Turner A, Millward GE. 2000. Particle dynamics and trace metal reactivity in estuarine plumes. Estuarine, Coastal and Shelf Science. 50:761-774. Ure A. 1991. Trace element speciation in soils, extracts and solutions. Mikrochimica acta. II:49-57. Ure A, Quevauvellir PH, Muntau H., Griepink B. 1993. Speciation of heavy metals in soils and sediments, an account of the improvement and harmonization of extraction techniques under- taken the auspices of the BCR of the commission of the European Communities. International Journal of Environmental and Analytical Chemistry. 51:135-151. USGS, GEOBOL, 1992. Geology and mineral resources of the Altiplano and Cordillera Occi- dental, Bolivia. U.S. Geological Survey Bulletin 1975, United States Geological Survey, & Ser- vicio Geológico de Bolivia, 365p. Varejao EVV, Bellato CR, Fontes MPF, Mello JWV. 2011. Arsenic and trace metals in river water and sediments from the southeast portion of the Iron Quadrangle, Brazil. Environmental Moni- toring and Assessment. 172:631-642.

Oswaldo Eduardo Ramos Ramos TRITA LWR PHD-2014:04

Walkley A. 1946. A critical examination of a rapid method for determining organic carbon in soils effect of variations in digestion conditions and of inorganic soil constituents. Soil. Science. 63:251-263. Wedepohl KH. 1995. The composition of the continental crust. Geochimica et Cosmochimica Acta. 59(7):1217-1232. Wenzel WW, Kirchbaumer N, Prohaska T, Stingeder G, Lombi E, Adriano T. 2001. Arsenic fractionation in soils using an improved sequential extraction procedure. Analytica Chimica Ac- ta. 436:309-323. Williams P, Villada A, Deacon C, Raab A, Figuerola J, Green A, Feldmann J, Meharg A. 2007. Greatly enhances arsenic shoot assimilation in rice leads to elevated grain levels compared to wheat and barley. Environmental Science and Technology. 41:6854-6859. Whiting KS, 1992. The thermodynamics and geochemistry of As with application to subsurface waters at the Sharon Steel Superfund Site Midvale, Utah. MSc thesis Colorado School of Mines. WHO, 2011. Guideline for drinking water quality, fourth ed. World Health Organization; Singa- pore. Geneva YPFB, GEOBOL. 1996. Mapa de Bolivia, Scale 1:1,000,000 with accompanying memorandum. Yacimientos Petroliferos Fiscales Bolivianos and Servicio Geológico de Bolivia. 27p. Zapata R, 2011. Hydrogeology concept model of aquifer system of an alluvial fan in the Poopó river sub-basin: Oruro-Bolivia. (Modelo conceptual hidrogeológico del sistema acuífero de un abanico aluvial en la sub-cuenca del río Poopó: Oruro-Bolivia). Universidad Mayor, Real y Pontificia de San Francisco Xavier de Chuquisaca, MSc thesis. 207p. Zheng R, Li H, Jiang R, Römheld V, Zhang F, Zhao F. 2011. The role of roots hairs in cadmium acquisitions by barley. Environmental Pollution. 159:408-415.

Other publications La Patria newspaper, 2010a. Increasing prices, royalties and usefulness, Perspectiva Minera. La Patria newspaper, 2010b. Increasing the activities in the national mining companies. Perspec- tiva Minera. CIDSE, 2009. Impacts of extractive industries in Latin America. URL:http//www.cidse.org. Perales V. H. 2010, Conflictos geolopoliticos por el agua en las cuencas mineras del Departamen- to de Oruro, Bolivia. VertigO, serie 7. URL:http//www.vertigo.revues.org/9769. Ministerio Metalurgia y Mineria. (MMM), 2013. (URL: http://www.mineria.gob.bo/Docu- mentos/Estadisticas/Exportaciones.pdf (visit date: 2013-04-15). Instituto nacional de estadística (INE). URL: www.ine.gob.bo (visit date 2014-03-20). El Diario newspaper, 2012. Bolivia: de los barones del estaño a los dueños de la minería. URL: http://eju.tv/2012/06/bolivia-de-los-barones-del-estaño-a-los-dueños-de-la-minera/ (visit date: 2014-03-20). DIMA-COMIBOL, 2014. Mining properties map. (URL: http://dimacomibol.gob.bo/en/do- nde_operamos/mapa_geografico. (visit date: 2014-04-28).