Clemens Reimann Manfred Birke (eds.) sample pages. Borntraeger Science Publishers

Transcription

1 Clemens Reimann Manfred Birke (eds.) Geochemistry of European Bottled Water Borntraeger Science Publishers

2 Geochemistry of European Bottled Water Edited by: Clemens Reimann Geological Survey of Norway (NGU) EuroGeoSurveys Geochemistry Expert Group, Norway and Manfred Birke Federal Institute for Geosciences and Natural Resources (BGR), Germany Back cover 1: The growth of the bottled water market in Germany, Hungary, Italy and Sweden from , see p. 7. 2: A well-dressing at the Derbyshire village of Barlow, England see p. 9. 3: ICP-QMS 7500 used for several analyses, 2010 Agilent Technologies, Inc. ISBN Information on this title: Gebr. Borntraeger Verlagsbuchhandlung, Stuttgart, Germany All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical photocopying, recording, or otherwise, without the prior written permission of Gebr. Borntraeger Verlagsbuchhandlung. Bottled waters (=groundwaters) are natural substances and as such subject to numerous infl uences. Therefore their compositions may be subject to change with time. The water compositional data published in this book were determined with utmost care, using state-of-the-art analytical equipment and refl ect the water compositions at the time of analysis. Publisher: Gebr. Borntraeger Verlagsbuchhandlung Johannesstr. 3A, Stuttgart, Germany mail@borntraeger-cramer.de Printed on permanent paper conforming to ISO Typesetting DTP + TEXT Eva Burri, Stuttgart, Germany Printed in Germany by Gulde Druck GmbH, Tübingen, Germany

3 III Table of Contents Project Staff... V Foreword by the Secretary General of EuroGeoSurveys... IX Foreword by the President of the Federal Institute for Geosciences and Natural Resources, Germany and the Director of the Geological Survey of Norway... XI Abstract Introduction The Concept of the Geochemistry of European Bottled Water Project What is Groundwater? What is Mineral Water? From Mineral Water Springs and Spas to the Bottled Water Market The Legal Framework: Bottled Waters, Spring Waters and Mineral Waters Natural Mineral Water Spring Water Other Bottled Water / Table Water Medicinal Waters What does this Atlas Represent? The Hydrochemistry of Groundwater The Infl uence of Rainfall The Infl uence of Vegetation The Infl uence of Soil Acid-Base Reactions Dissolution Reactions Redox Reactions Mixing with deep Saline Waters Carbon Dioxide Water Types Background Information Topography Geology Mineral Deposits of Europe Hydrogeological map of Europe Soil Climate Vegetation and Land Use Human Activities Methods Sampling Analyses Sample preparation ICP-QMS analyses ICP-AES analyses Atomic fl uorescence spectrometry (AFS) Ion Chromatography (IC) Photometric analyses Titration Potentiometry Conductometry... 35

4 IV Table of Contents 4.3 Quality Control International Reference Materials The in-house MinWas project standard Sample blanks Detection limits Duplicate analyses Comparison of analytical results with the water composition provided on the labels Re-analyses of selected samples Possible Problems with the Data in Terms of Representing True Groundwater Quality or being able to Refl ect Geology Bottle leaching The effect of carbonatisation Other forms of water treatment Further possible disturbing factors Data Analysis and Mapping Datasets Bottled Water European Tap Water Other Datasets Comparing Bottled Water with European Tap Water, Hardrock Groundwater from Norway and European Surface Water Maps and their Interpretation Health Implications Elements with established MAC-Values for Mineral Water (EU Directive 2003/40/EC Mineral Water) Antimony, p. 202; Arsenic, p. 202; Barium, p. 203; Cadmium, p. 203; Chromium, p. 203; Copper, p. 203; Fluoride, p. 204; Lead, p. 204; Manganese, p. 204; Mercury, p. 204; Nickel, p. 205; Nitrate, p. 205; Nitrite, p. 205; Selenium, p Elements (Parameters) with Established MAC (or Guideline)-Values for Drinking Water (EC Directive 98/83/EC) but not for Mineral Water Aluminium, p. 205; Ammonium, p. 206; Boron, p. 206; Chloride, p. 206; Electrical conductivity (EC), p. 206; Iron, p. 206; ph-value, p. 206; Sodium, p. 206; Sulphate, p Elements with non-eu Drinking Water MAC Values Silver, p. 207; Beryllium, p. 207; Molybdenum, p. 207; Strontium, p. 207; Thallium, p. 207; Uranium, p. 207; Vanadium, p. 208; Zinc, p Other Health-Relevant Elements Iodine, p. 208; Lithium (Rubidium, Caesium), p. 208; Rare Earth Elements, p. 209; Thorium, p Additional Remarks Conclusions References Appendix A (Table of Water Action Levels as defi ned by the EU and for the majority of Countries from which water samples are included in this study) Appendix B (Statistical summary table for the datasets used in this Atlas) Contents included on the Data CD-ROM

5 V Project Staff Project management: Manfred Birke, Federal Institute for Geosciences and Natural Resources (BGR), Germany Clemens Reimann, Geological Survey of Norway (NGU), Chairman EuroGeoSurveys Geochemistry Expert Group, Norway Preparation of Atlas Manuscript: Clemens Reimann, NGU David Banks, Holymoor Consultancy Ltd., Chesterfi eld, U.K. Manfred Birke, BGR Alecos Demetriades, IGME Analyses: Manfred Birke, Hans Lorenz, Wolfgang Glatte, Bodo Harazim, Fred Flohr, Anna Degtjarev, Jürgen Rausch, Federal Institute for Geosciences and Natural Resources, Germany Statistics and R-Scripts: Peter Filzmoser, Institute of Statistics and Probability Theory, Vienna University of Technology, Austria Geology: Fabian Jähne, Federal Institute for Geosciences and Natural Resources, Germany General information maps: Ola Eggen, NGU; Marianthi Stefouli, IGME (digital processing of Global Land cover map), Uwe Rauch, BGR (topography) Hydrogeology: David Banks, Holymoor Consultancy Ltd., Chesterfi eld; Christophe Innocent, BRGM; Bjørn Frengstad, NGU, Carla Lourenço, LNEG ex-ineti; Pauline Smedley, BGS With input (sampling and interpretation) from the following Country Teams: Albania Friedrich Koller, University of Vienna (sampling) Kujtim Onuzi, Institute of Geoscience, Technical University of Tirana Austria Gerhard Hobiger, Albert Schedl, Geological Survey of Austria (GBA) Edith Haslinger, Austrian Institute of Technology (AIT) Peter Filzmoser, Vienna University of Technology Friedrich Koller, University of Vienna Belarus Peter Filzmoser, Vienna University of Technology (sampling) Virgilija Gregorauskiene, Lithuanian Geological Survey (sampling) Belgium Walter De Vos, Geological Survey of Belgium (interpretation) Ilse Schoeters, Rio Tinto Minerals (sampling) Clemens Reimann, NGU (sampling)

6 VI Project Staff Bosnia & Herzegovina Hazim Hrvatovic, Neven Miosic, Ferid Skopljak, Natalija Samardzic, Geological Survey of the Federation of Bosnia and Herzegovina Bulgaria Valeri Trendavilov, Dept. of Geology and Permits for Exploration, Subsurface and Underground Resources Office, Ministry of Environment and Water Croatia Josip Halamić, Ajka Šorša, Croatian Geological Survey Czech Republic Miloslav Duris, Czech Geological Survey Cyprus Alecos Demetriades, Hellas for Cyprus Geological Survey Department Denmark Rolf Tore Ottesen and Clemens Reimann, NGU (sampling) Estonia Jaan Kivisilla, Valter Petersell, Geological Survey of Estonia Liida Bityukova, Institute of Geology at Tallinn University of Technology, Estonia Finland Timo Tarvainen, Jaana Jarva, Geological Survey of Finland Former Yugoslavian Republic of Macedonia (F.Y.R.O.M.) Trajce Stafilov, Sts. Cyril and Methodius University, Skopje France Ignace Salpeteur, Christophe Innocent, BRGM Germany Manfred Birke, Uwe Rauch, Lars Kaste, Hans Lorenz, Federal Institute for Geosciences and Natural Resources (BGR), Germany Hellas Alecos Demetriades, Institute of Geology and Mineral Exploration (IGME) Hungary Gyozo Jordan, Ubul Fugedi, Laszlo Kuti, Geological Institute of Hungary (MAFI) Iceland Børge Johannes Wigum, Mannvit hf Ireland Raymond Flynn, Queen s University Belfast, Northern Ireland Manfred Birke, BGR (sampling) Italy Benedetto De Vivo, Annamaria Lima, Stefano Albanese, University of Napoli Federico II; Domenico Cicchella, University of Sannio; Enrico Dinelli, University of Bologna; Paolo Valera, University of Cagliari Latvia Aivars Gilucis, Latvian Environment, Geology and Meteorology Agency Lithuania Virgilija Gregorauskiene, Lithuanian Geological Survey Luxembourg Robert Maquil, Service Géologique du Luxembourg (SGL)

7 Project Staff VII Moldova Maria Titovet, Ministry for Ecology and Nature of Republic Moldova, Department of Hydrogeology Montenegro Neda Devic, Geological Survey of Montenegro Netherlands, The Rolf Tore Ottesen, Clemens Reimann, NGU (sampling) Norway Bjørn Frengstad, Rolf Tore Ottesen, Clemens Reimann, Geological Survey of Norway (NGU) Poland Lech Smietanski, Polish Geological Institute, Warsaw Portugal Carla Lourenço, Maria Joao Batista, Laboratório Nacional de Energia e Geologia, I. P. (LNEG ex-ineti) Romania Adriana Ion, Geological Institute of Romania Corina Ionesco, University of Cluj-Napoca Russia Nikolay Phillipov, SC Mineral Olga Karnachuk, Tomsk State University Reijo Salminen, Geological Survey of Finland (sampling) Serbia Milena Zlokolica-Mandic, Tanja Petrovic, Aleksandra Gulan, Geological Institute of Serbia; Nebojsa Veljkovic, Serbian Environmental Protection Agency Slovakia Peter Malik, State Geological Institute of Dionyz Stur, Bratislava Slovenia Mateja Gosar, Geological Survey of Slovenia Spain Juan Locutura, Alejandro Bel-lan, Mar Corral, Instituto Geologico y Minero de Espana (IGME). Sweden Kaj Lax, Madelen Andersson, Geological Survey of Sweden (SGU) Switzerland Peter Hayoz, Federal Offi ce of Topography, Swiss Geological Survey Turkey Manfred Birke, BGR (sampling), Clemens Reimann, NGU (sampling) Ukraine Boris I. Malyuk, Volodymyr Klos, Maryna Vladymyrova, Geological Survey of Ukraine United Kingdom Dee Flight, Shaun Reeder, Pauline Smedley, British Geological Survey David Banks, Holymoor Consultancy Ltd.

8

9 IX Foreword by the Secretary General of EuroGeoSurveys EuroGeoSurveys (EGS) represents 32 European Geological Surveys with a staff of over 20,000 specialists working in numerous applications of geosciences for EU society and economy. EuroGeoSurveys is a not-for-profi t association working solely for the public interest. EGS aims at providing the European Institutions with expert, neutral, balanced and practical pan-european advice and information as an aid to problem-solving, policy, regulatory and programme formulation in areas such as: the use and the management of on- and off-shore natural resources related to the subsurface of the Earth, (energy, including the renewable geothermal energy; minerals and water, soils, underground space and land); the identification of natural hazards of geological origin, their monitoring and the mitigation of their impacts (depletion or excess of trace elements in soils and waters, earthquakes, natural emissions of hazardous gases, landslides and rockfalls, land heave and subsidence, shrinking and swelling clays; environmental management, waste management and disposal; land-use planning; sustainable urban development and safe construction; e-government and the access to geoscientifi c metadata and data; the development of interoperable and harmonised geoscientifi c data at the European scale. The portal of EGS ( provides access to geoscientifi c metadata, information and knowledge at European and national scales, following the links on the thematic pages. It also presents information on EuroGeoSurveys its activities and Member Organisations. Geology knows no borders. However, up to date most investigations of the national Geological Surveys stop at political country borders. Yet, at the same time the administration of the EU urgently requires for its policy directly comparable data and information at the European scale, an overview for the whole continent. For some issues, overview maps at the continental scale already exist; examples are the Geological Map of Europe (Asch, 2003), the Hydrogeological Map of Europe (IHME1500, 2008), the Soil Map of Europe (Jones et al., 2005) or the results of the Corine Land Cover project (GLC2000 database, 2003). From 1996 to 2006 twenty-six European Geological Surveys cooperated to produce the Geochemical Atlas of Europe (Salminen et al., 2005; De Vos et al., 2006), presenting for the fi rst time directly comparable data of the chemical composition of topand sub-soil, stream water, stream sediment and floodplain sediment at the European scale. With the publication of the two atlas volumes, the datasets were made publicly available and are now used in a wide variety of applications by diverse organisations throughout Europe, including European Commission institutions. Although each country has its own national database on groundwater quality with which it makes management and regulatory decisions, a directly comparable dataset on natural groundwater quality at the European scale is not available. A pan-european harmonised dataset of natural groundwater quality would contribute signifi cantly to an understanding of groundwater quality across Europe and may support implementation of the EU Water Framework Directive. Collecting representative groundwater samples at the European scale is not an easy task (see discussion in this atlas), and may be prohibitively expensive, if done at a high sample density. It was because of this that one of the members of the Euro- GeoSurveys Geochemistry Expert Group, Dr. Manfred Birke, put forward the idea that groundwater can be sampled by purchasing bottled waters from supermarkets over all of Europe, and suggested to use the EuroGeoSurveys network to collect bottled water as a fi rst proxy for groundwater. Though the idea met some resistance, it was in the end decided that it was worth a try. The present atlas is the result of this rather unusual, and somewhat innovative, project. This volume could be considered as the third in the EuroGeoSurveys series of Geochemical Atlases of Europe. Results show the enormous variation of many chemical elements in bottled water derived from groundwater at the European scale, usually 3 to 4 and up to 7 orders of magnitude. The importance of geology in controlling natural water chemistry is clear a fact that needs

10 X Foreword by the Secretary General of EuroGeoSurveys to be considered in all discussions about contamination and when managing groundwater. The analytical results are made available together with this atlas, and we hope that they will be of general interest across Europe and, may be for research and teaching purposes at universities. Until now, there has not been a groundwater dataset covering such a large area in Europe where all samples were analysed in just one laboratory and which delivers directly comparable results for so many parameters. Luca Demicheli EuroGeoSurveys Secretary General EuroGeoSurveys Offi ce Rue Joseph II Brussels, Belgium Tel : ; Fax : luca.demicheli@eurogeosurveys.org info@eurogeosurveys.org URL:

11 XI Foreword by the President of the Federal Institute for Geosciences and Natural Resources, Germany and the Director of the Geological Survey of Norway The Science of geochemistry aims to quantitatively determine the chemical composition of the Earth and its parts and to discover the factors that control the distribution of the individual elements (Goldschmidt, 1937, 1954). Applied geochemistry is the application of this knowledge to societal benefi t, whether for discovering mineral resources, protecting the surface environment that sustains life, improving the effi ciency of agriculture and animal husbandry, or studying the behaviour of elements in the food-chain and their health effects on humans (epidemiology) and other biota. Traditionally geochemical surveys have been carried out at a very high density, based on the argument that natural variation is very high even in small areas. Only during the last decade, has low-density geochemistry (1 sample site per several hundred to ten thousand km 2 ) covering whole continents seen a breakthrough (Garrett et al., 2008; Smith and Reimann, 2008), delivering much needed information about the processes dominating the geochemical distribution of elements at the Earth s surface at the continental scale. Working at the continental scale is still not a trivial task in Europe; geochemical surveys are most often carried out at the national scale and are not comparable across political borders. At the same time the political landscape in Europe is rapidly changing and databases that are directly comparable at the European scale are needed for important political decisions. High quality geochemical databases are an essential component of environmental knowledge; they provide pertinent information for administrative and legal issues. Sustainable management of European resources depends on reliable databases like those provided in the Geochemical Atlas of Europe series (Salminen et al., 2005; De Vos et al., 2006) and continued with this volume. The statistics and maps provided in this book are stimulating and prove the importance of geology in controlling the natural chemical composition of groundwater. For many elements the natural variation in element concentration is enormous and surprisingly clear distribution patterns emerge that can be directly linked to certain bedrock types or geological processes. Morten Smelror, Director, Geological Survey of Norway Trondheim Hans-Joachim Kümpel, President, Federal Institute for Geosciences and Natural Resources, Hannover

12

13 1 Abstract As of the start of the year 2010, 1916 mineral water brands are offi cially registered in Europe. Bottled water (usually derived from groundwater) is rapidly developing into the main drinking water supply for the general population in large parts of the continent. Analysis of bottled water may thus provide a cheap possibility to gain an idea about groundwater chemistry at the European scale. For this study, 1785 bottled water samples were purchased, representing 1247 wells at 884 locations and analyzed for more than 70 parameters. The dataset is used to gain a first impression about the natural concentrations of, and variations in, the determined chemical elements and additional parameters in groundwater at a European scale. Many processes affect the hydrochemical fi n- gerprint of groundwater important factors include: rainfall chemistry, climate, vegetation and soil zone processes, mineral-water interactions, groundwater residence time and the mineralogy and chemistry of the aquifer (and contamination). The infl uence of geology in determining element concentrations in bottled water can be observed for a signifi cant number of elements. Examples include: high values of chromium (Cr) clearly related to the occurrence of ophiolites; beryllium (Be), caesium (Cs), germanium (Ge), potassium (K), lithium (Li) and rubidium (Rb) showing unusually high values in areas underlain by Hercynian granites while high values of aluminium (Al), arsenic (As), fl uorine (F), potassium (K), rubidium (Rb) and silicon (Si) in the bottled water are related to the occurrence of alkaline volcanic rocks. A further key observation is that knowledge of geology alone is inadequate to predict the hydrochemistry of bottled water: natural variation is enormous, usually three to four and for some elements up to seven orders of magnitude. Such variation may refl ect, among other factors, groundwater residence time and mixing with deep brackish formation waters. It has also been found that bottle materials can have an infl uence on bottled water chemistry. For antimony (Sb), leaching from the bottle material is so serious that the results for bottled water cannot be used as an indication of natural concentrations in groundwater. Some elements, as observed in the bottled water, are clearly not representative of typical, shallow, fresh groundwater; rather, they tend to exhibit unusually high concentrations, typical for mineral water : examples are boron (B), beryllium (Be), bromine (Br), caesium (Cs), fl uorine (F), germanium (Ge), lithium (Li), rubidium (Rb), tellurium (Te), and zirconium (Zr). Very few analysed samples (in general less than 1%) returned values exceeding maximum admissible concentrations (MACs) for mineral water, as defi ned by the European Commission.

14

15 3 1 Introduction 1.1 The Concept of the Geochemistry of European Bottled Water Project In 2005 the Forum of European Geological Surveys (FOREGS), now EuroGeoSurveys (EGS), published the Geochemical Atlas of Europe (Salminen et al., 2005), followed in 2006 by an interpretation volume (De Vos et al., 2006). These volumes presented for the first time directly comparable chemical analyses (more than 50 elements) in fi ve principal sample materials (surface (stream) water, stream and floodplain sediments, top- and sub-soil) at the European scale. The Atlas clearly demonstrated that the geochemical patterns observed at a continental scale are, contrary to common opinion, almost completely governed by natural sources and processes. One of the great surprises observed was a major discontinuity in concentration levels for the majority of chemical elements along the border of the last glacial maximum, with much lower element concentrations (often by a factor of ca. 3) observed in most sample materials and for the majority of elements in Northern Europe when compared to Southern Europe. Thus at least two strongly different geochemical background regimes (Reimann and Garrett, 2005) are observed at the European scale. Comparable geochemical background data at the European scale are urgently needed by regulators for setting action levels or for judging the impact of contamination on the natural environment. In the FOREGS atlas (Salminen et al., 2005; De Vos et al., 2006) a missing, but rather important, sample material is groundwater. Groundwater is a diffi cult sample material for geochemical mapping projects for a variety of reasons: (1) The location of wells is irregular, depending on where people live and population density and permeability of underlying aquifers geochemical maps with a high sample density in small areas and large gaps in between will never be aesthetically pleasing and the data are often not suitable for the construction of smoothed surface maps (Reimann, 2003). (2) Different groundwater chemistry may be observed at different depths at one and the same site, depending on well depth (several aquifers may exist at one and the same location). For example groundwater may be abstracted from shallow springs or deep boreholes; from shallow Quaternary glaciofl uvial sedimentary aquifers or from deep crystalline hard rock aquifers, at one and the same location. (3) The residence time of the groundwater can differ substantially from well to well. Even the water abstracted from a single well can represent a mixture of groundwaters with differing residence times, derived from different flow horizons or fractures. (4) Groundwater chemistry can vary seasonally and from year to year. (5) Water wells can be differently constructed, according to local culture and perceived best practice; the groundwater may thus come into contact with different construction materials, water pipes and installations. It is therefore almost inevitable that the water will be, in some sense, contaminated already before sampling. (6) Access to many wells can often be quite diffi cult. Many of the above difficulties fi appear to have no ob- vious solution, and one might be tempted to give up on the idea of constructing a comparable groundwater geochemistry database at the European scale. Such a database would, however, be highly desirable and is actually required, for example, in connection with the European Water Framework Directive (EU directive 2000/60/EC) and the new groundwater directive (EU directive 2008/32/EC). At present, natural variation in groundwater quality is not well understood due to lack of compatible groundwater data across Europe. Water legislation is dominantly focussed on regulating contamination, while natural variation and natural risks may not always be suffi ciently recognized. Though many surveys of groundwater quality exist at the country scale (see Edmunds and Shand, 2008 for a general overview and: Edmunds et al., 1989; Lahermo et al., 1990; Schleyer and Kerndorff, 1992; Aastrup et al., 1995; Rapant et al., 1996; Frengstad et al., 2000; Tarvainen et al., 2001; Serra et al., 2003; Roux and Pointet, 2005; Brenot et al., 2007; Shand et al., 2007 for examples of detailed studies), good and directly comparable data at the European scale are missing. It is a central task of Geological Surveys to document the natural variation of chemical elements in a wide selection of sample materials, starting with bedrock, but including stream and fl oodplain sedisample pages

16 4 Introduction ments, marine sediments, surface and groundwater, soils and even vegetation. At a meeting of the EGS Geochemistry Expert Group, it was suggested that it might be very easy and relatively cheap to obtain a first overview of European Groundwater Geochemistry, element concentrations and variation at the European scale, by purchasing and analysing commercially available bottled water (i.e. mineral water and spring water). Almost 2000 (1916) natural mineral water brands are named in the official fi list recognized by the Member States of the European Union (see eu/food/food/labellingnutrition/water/mw_eulist_ en.pdf last accessed: ). During the past 20 years it has become highly fashionable to drink bottled water and never before have so many groundwater sources (wells and springs) been in production and on the market. Hence groundwater as drunk can be easily sampled because it is sold pre-packaged and analysis-ready in supermarkets all over Europe. The EGS Geochemistry Expert Group has representatives in all European countries. It was decided to use this network to buy as many different water samples as possible throughout Europe and to ship all samples to one central laboratory, equipped and capable to analyze these samples with the required low detection limits for many of the trace elements. There exist already numerous publications on the chemical composition of mineral (bottled) water (e.g., Misund et al., 1999; Dabecka et al., 2002; Lau and Luk, 2002; Van der Aa, 2003; Rosborg et al., 2005; Karamanis et al., 2007; Shotyk et al., 2006; Soupioni et al., 2006; Shotyk and Krachler, 2007a,b; Baba et al., 2008; Naddeo et al., 2008; Güler and Alpaslan, 2009; Krachler and Shotyk, 2009; Bertoldi et al., 2010), but the focus of the vast majority of these studies was on MACs and health issues and not on the importance of geology for the observed element concentrations. Furthermore none of these studies presents data for as many samples analyzed for as many parameters as provided here. Of course, a number of additional problems were subsequently identified when planning to use water sold in a variety of different bottle types for ultra trace element analysis: (1) Bottled water may be treated (though only rather limited treatment is allowed by law for spring and mineral water); for example, carbon dioxide may be added (or removed), certain unstable elements such as iron, manganese, arsenic and sulphur may be removed (and, with the removal of these, a range of minor elements may be scavenged from the water). Water from several sources may, occasionally, be mixed to dilute the concentrations of certain critical elements. Treatment should be, but is not necessarily, indicated on the labels. (2) Geochemists are well aware that bottle material may contaminate the water stored in the bottles (e.g. Hall, 1998; Reimann et al., 1999, 2007). Contamination may differ from bottle type to bottle type European bottled waters are sold in soft polyethylene terephthalate (PET) bottles, in hard PET bottles, in glass bottles and even in Al-containers or tetrapacks. All these bottles are in addition marketed in a variety of colours and sizes. (3) Mineral and spring water must be produced from a protected source (and such sources are often relatively deep) and may thus not refl ect the real extent of contamination, especially of shallow groundwater aquifers. (4) Mineral water may quite often be special water ; for example, geothermal water (see Serra et al., 2003) or water abstracted at great depth. It could thus be unusually salty in comparison to normal shallow groundwater. The preference for such salty bottled water may even be culturally determined. Again, all these problems can be used as strong arguments against the approach of using bottled water as a proxy for European groundwater quality. However, it is difficult fi to evaluate beforehand how seriously any of these potential problems will affect the variation in water chemistry at a European scale. Some of the effects can be tested beforehand (e.g. the influence of different bottle types see Chapter on Bottle leaching ). Moreover, the bottled water can be compared to European surface water (from Salminen et al., 2005) or tap water, representing drinking water, collected at the European scale. On balance, the EGS Geochemistry Expert Group decided that the approach was potentially valuable and informative: it proceeded to collect and analyze as many of the bottled waters on the European market as possible, in an attempt to assess the variation of bottled water quality across Europe and its relationship to geology. Groundwater is the water that exists in pore spaces, fractures, fi ssures and caverns in the rocks and sediments beneath our feet. Below the groundwater table all voids are entirely fi lled with groundwater while above, in the so-called vadose zone, cavities are partly fi lled with air. Groundwater is usually ultimately derived from rainfall soaking through the soils and then into the rocks, although sometimes it may have its origin in sea water or river water seeping into an aquifer. During its passage through a wasample pages This atlas presents the results, maps the data and evaluates whether the attempt was justified. 1.2 What is Groundwater?

17 What does this Atlas Represent? 11 be of such high microbiological quality that no disinfection is required. Its composition and temperature must remain stable within the limits of natural fl uctuation and must not vary with fl ow rate. The water must not be treated or disinfected at all, except for the limited application of simple processes such as (ozone-enriched) aeration, decanting and fi ltration to remove certain unstable elements such as iron (Fe), manganese (Mn), arsenic (As) and sulphur (S). The carbon dioxide content of a natural mineral water may be manipulated, and the sold product can be designated as one of the following: Uncarbonated, Carbonated, Naturally carbonated, Fully decarbonated or Partially decarbonated natural mineral water or as natural mineral water fortifi ed with gas from the spring. A natural mineral water source must have been fully characterized in terms of its geology and chemistry and should be protected against pollution. A single natural mineral water source cannot supply more than one mineral water product. The source and the water composition must be clearly identifi ed on the label. Natural mineral waters do not necessarily need to satisfy normal drinking water quality requirements: they need only satisfy the limits established by EC Directive 2003/40/EC. All the maximum admissible concentrations refer to natural contents: the contents should not be the result of contamination at source Spring Water According to Directive 96/70/EC, the term spring water shall be reserved for a groundwater which is intended for human consumption in its natural state, and bottled at source, which satisfi es many of the same conditions as for mineral waters regarding microbiological purity, source protection, labelling and treatment. However, the term spring water does not imply the same requirement for a characteristic mineral composition, or for the same degree of stability of composition. Finally, the requirements for source characterization are not as strict as for natural mineral waters and there may be slightly more scope for treatment of the water. However, spring waters must satisfy the normal limits for the quality of drinking water (EU Directive 80/778/EEC, subsequently replaced by 98/83/EC), which may be more stringent than the Mineral Water limits (see Table of European quality constraints in the different water types in the Appendix) Other Bottled Water / Table Water Products that are marketed simply as bottled water or table water are not subject to any specifi c requirements other than that they must satisfy the normal limits for the quality of drinking water. They can simply be normal tap water or treated and disinfected river water Medicinal Waters The final, and rather unusual, category of bottled water is that of Medicinal Water. The criteria for this designation are rather stringent and are governed by EU Directive 65/65/EEC. In brief, this Directive effectively subjects Medicinal Waters to the same testing and labelling regimes as other pharmacological products that claim to have a health effect. In this connection it may appear surprising that several medicinal waters can be purchased in the supermarket and from the same shelves as non medicinal bottled water in many EU countries. 1.5 What does this Atlas Represent? This Atlas documents the quality of bottled spring and mineral waters across Europe, based on an unevenly geographically distributed dataset of samples purchased on the available market in 2008/2009. It provides an excellent starting point for answering some of the concerns raised towards the end of the previous section, with regards to the possible health impacts of basing a diet on the consumption of bottled water. But, is it possible to go any further? Can one dare to hope that this Atlas is in some way representative of European groundwater quality? Certainly, the patterns seen on the maps for some ele- ments do indeed seem to refl ect large scale natural features, where water chemistry is systematically affected by regional climate, geology or other factors such as land use or vegetation cover as, for example, the maps for As, Cr and V clearly demonstrate. Moreover, the authors of this Atlas believe that the range of element concentrations represented in this Atlas (from the very dilute bottled waters of Norway to the mineral-rich waters of Eastern Europe) is reasonably representative of the range of elements naturally present in European groundwater (though for some specifi c elements, the mineral waters may represent the high end of this range). The Atlas is thus an excellent starting point for gaining some impression of European groundwater chemistry. If one looks with a rigorous sceptic s eye at the dataset, however, there are good reasons why one should be reluctant to claim that the bottled water dataset is truly statistically representative (in terms of intricacies of means, medians and percentile distributions!) of overall groundwater quality for all the elements presented. These reasons are mentioned at several occasions throughout this atlas but

18 12 Introduction are summarized here once more in small print to clearly state the limitations of the chosen approach. Some of the reasons for possible non-representativity are obvious: Bottled spring and mineral water is, by defi nition (one would hope!), protected from point source pollution and from the worst excesses of diffuse pollution (although a glance at the map for nitrate would suggest that the diffuse impact of intensive agriculture can be discerned even in sources used for bottling). The bottled water dataset thus tells very little about the impact of human activity on groundwater. Although legislation restricts permitted treatment of spring and mineral waters, some treatments are allowed. Thus, the iron (Fe), manganese (Mn) and arsenic (As) contents of bottled waters may be depleted relative to the original groundwater, due to treatment by aeration and settlement of iron and manganese oxyhydroxides. Some waters may have been legitimately diluted to reduce elevated concentrations of uranium. Some bottled water brands have been artifi cially gasifi ed with carbon dioxide. This should not signifi - cantly alter the elemental composition of the water, but will affect its ph and alkalinity significantly (Figures 21 and 24). The bottle material itself may have an influence on water quality. For example, this study has clearly shown that glass bottles have a tendency to leach lead (Pb) into the water, while many plastic (PET polyethylene terephthalate) flasks have a tendency to leach antimony (Sb). It should be stressed that the quantities leached are well below health-related drinking water standards, but the effect is signifi cant in the context of natural groundwater chemistry (Figure 22). While other reasons are more subtle: The very idea of a truly representative European groundwater dataset is highly elusive. What does representative mean when one is talking about a three-dimensional, dynamic medium? Groundwater quality varies with time and depth how could one ever hope to pin it down on a 2-dimensional map? At any given point on a map, for example, the chemistry of deep groundwater may be very different from that of shallow groundwater. It has been mentioned that, in some cultures, mineral water will be attractive because it is atypical for example, because consumers prize its very high mineral content or its dissolved gas content (true mineral water, which continues to be popular in Central and Eastern Europe). Alternatively, bottled water may be atypical because of its purity or its low mineral content (many Western European bottled waters). The maps for some elements in bottled water may refl ect cultural preferences rather than natural variation: this can be very clearly seen on the map for sodium (Na), where the salt content of bottled water increases strongly to the South and East of Europe. This is opposite to what one would expect intuitively from geography (i.e. highest salt contents near the open Atlantic), and is possibly due to a strong cultural preference for mineral-rich waters in Eastern Europe (but may also be due to the deep sedimentary basins in these regions and thus quite representative of geology ). However, it is possible to compare the results of this study (bottled water) to other studies on groundwater or general water quality at a European or a more local scale. To answer the most obvious question, How does the European bottled water compare to European tap water?, no suitable dataset was available. The EGS Geochemistry Expert Group thus decided to collect a number of tap water samples from all over Europe (n=579), again utilizing the network of colleagues and friends, and to analyze these tap water samples in the same laboratory that carried out the bottled water analyses (see below for details). Furthermore, the European surface water data presented in Salminen et al. (2005), as well as a dataset from Norwegian crystalline bedrock groundwater wells (Banks et al., 1998; Frengstad et al., 2000), can be used for comparison with the bottled water. For all these datasets almost the entire suite of elements reported for the bottled water was analyzed.

19 13 2 The Hydrochemistry of Groundwater 2.1 The Influence of Rainfall Groundwater typically begins its hydrological journey as rainfall, which falls from the sky and seeps into the ground. Rainfall may be distilled water, but it is not pure water. Even at this stage it contains some dissolved chemicals, or solutes. For example, it contains dissolved oxygen (O 2 ), nitrogen (N 2 ), carbon dioxide (CO 2 ) and other gases that it has picked up from the atmosphere. Even if the world had been untouched by human activity, the ph of rainfall would be slightly acidic, somewhere around ph = 5.2 to 5.6, because carbon dioxide forms a weakly acidic solution when it is dissolved (carbonic acid). However, for hundreds of years, human activities have been emitting gases such as nitrogen oxides, sulphur dioxide and hydrogen chloride to the atmosphere. All of these dissolve in rainfall and make it even more acidic (hence the term acid rain, which is applied to contaminated rainfall over much of Europe). Moreover, the gases increase the content of nitrate, sulphate and chloride in rainfall by a few mg/l. HCl H + + Cl (1) Hydrogen chloride gas acid + chloride O 2 + 2SO 2 + 2H 2 O 4H + + 2SO 4 2 (2) Oxygen + sulphur dioxide acid + sulphate + water Storms at sea and waves along the coast will whip up salty sprays of sea water, which may become entrained by strong winds. These marine aerosols of sea spray may then become dissolved in rainfall near the coast, adding a content of marine salts (for example, sodium and chloride) to the rainfall. The content of sodium and chloride in rainwater and, hence, also in river water and groundwater, often decreases with increasing distance from the coast (see e.g. Reimann et al., 2000; Banks et al., 2001). Other processes may infl uence rainfall chemistry. Areas with large expanses of ploughed agricultural land or of desert may be vulnerable to strong winds blowing fi ne particles of soil or salt into the atmosphere, where they may become dissolved in rainfall. Alternatively, some industries, such as cement factories, may emit particles of basic oxides which will neutralize the acid rainwater in their surroundings. Thus, already when rainfall hits the ground, it contains up to several mg/l of inorganic solutes from natural and anthropogenic sources. 2.2 The Influence of Vegetation When the rain drops fall on leaves, they may partially evaporate. When they seep into the shallow soil they may be sucked up by the roots of plants and transpired to the atmosphere via the leaves. Such evapotranspirative processes return water to the atmosphere, but leave the dissolved salts behind. The net effect is to increase the concentration of many solutes in newly infiltrated rainfall, sometimes by a factor of three to four or more. A few types of solute are removed by plants and consumed as nutrients. Ions such as potassium (K + ), nitrate (NO 3 ) and phosphate (PO 4 3 ) fall into this category. The concentration of these may thus decrease as they pass through the vegetation and soil layer. 2.3 The Influence of Soil The soil usually contains a huge population of bacteria, fungi and other micro-organisms. These typically respire e they consume oxygen and release carbon dioxide. Thus, while the atmosphere contains about atm. of CO 2, the partial pressure of CO 2 in soil gas can be as high as atm. (100 times as high) and soil water is consequently somewhat more acidic than rainfall. The water that seeps out of the base of the soil zone therefore contains: higher contents of solutes than rainfall, due to evaporation strongly elevated contents of acidic CO 2 some residual atmospheric oxygen (O 2 ) Very hydrochemically young and immature groundwater may have a chemical composition that closely refl ects this recharge water (i.e. acidic ph, low solute content dominated by atmospheric and marine salts). This oxidizing, acidic recharge water now enters the geological environment, which is typically dominated by basic, reduced minerals. It is thus not

20 76 Arsenic Arsenic is a trace element with a predominantly non-metallic character (atomic number 33). It occurs in oxidation states -3, +3 and +5 in nature. Arsenopyrite (FeSAs) is a common As mineral. Many ore minerals (e.g. galena, pyrite, sphalerite) contain high concentrations of As, but also common silicates (like feldspars) can contain traces of As. Because of the possible substitution of P V by As V phosphate minerals can also contain As. Many As-minerals are relatively soluble, however, the mobility of As in the secondary environment is limited because of its strong tendency to sorb to clays, secondary hydroxides (especially ferric oxyhydroxides) and organic matter. Hydrothermal processes lead to an enrichment of As. It is, together with Sb, a typical element enriched in hydrothermal waters and an indicator of hydrothermal alteration zones (Banks et al., 2004). It is also often enriched in coal and black shales. Arsenic can be released to the groundwater environment as ferric oxyhydroxides (with a possible loading of sorbed arsenic) are converted by reductive dissolution to dissolved ferrous iron under chemically reducing conditions. This process is believed to be one of the dominant mechanisms for the evolution of arsenic-rich groundwater in some aquifers in Bangladesh. Conversely, when oxidation causes the precipitation of ferric hydroxides from solution, arsenic can also be removed from solution by sorption to the precipitate (this aeration process is one of the few permitted treatments of mineral / spring waters, in order to remove iron and manganese, and it would be expected to also remove arsenic from the water, if present). As-compounds are used in many herbicides, insecticides and fungicides, combustion processes, ore roasting and even pig and poultry farming can lead to an anthropogenic load of As. However, due to the strong redox- and ph-dependence of As solubility, the question of whether such contamination will reach groundwater will be very dependent on local conditions. The many incidences of high As in drinking water wells reported in literature (e.g. Smith et al., 2002; Smedley and Kinniburgh, 2002) are invariably connected to special redox- and geological conditions, are usually natural, and no contamination. The mean concentration of As in surface water has been estimated to be 4 μg/l (see Koljonen, 1992), a value that is likely to be too high in the light of the results for European surface water (0.63 μg/l Salminen et al., 2005) and the values presented here (0.24 μg/l) for bottled water. The map of As in bottled water shows quite a number of areas with enhanced As-values. Of the interesting patterns observed for As in the geochemical atlas of Europe were generally the higher concentrations in Southern than in Northern Europe (a factor 3 difference in the median value). The same can be seen for the bottled water, As values are clearly higher in the Southern European countries (see boxplots, median Northern Europe = 0.07, Southern Europe = 0.28 μg/l). Many of the high values displayed on the map occur in parts of Europe that are known for the occurrence of sulphide mineralization. Furthermore, the alkaline volcanic provinces in Italy are marked by higher than usual As values. The highest value (89.8 μg/l) was reported in a Polish mineral water, abstracted in an area with known mineralization. Even four samples from Germany are above the drinking water standards, one is abstracting the water from Palaeozoic rocks, another drawing water from a major fault zone. It is important to mention what is not seen on the map: in the Southern parts of the Great Hungarian Plain occurs groundwater with high natural concentrations of As (e.g. Smedley and Kinniburgh, 2005), which are of course not bottled, but would add on another order of magnitude to the natural variation observed for As. Bedrock groundwater with high As concentrations from Finland and Sweden are also not seen in this low density sample map. Bottle leaching should have no infl uence on the observed As concentrations (up to 0.09 μg/l from glass bottles). The water standard defined for As in mineral water and drinking water in general is 10 μg/l in the EU. Nine samples exceed this level. In the CP-diagram in Figure 26 all samples show very comparable As values and variation, with a slight enrichment of As in the European surface water in the lower and middle concentration range. The CP-diagram suggests that the natural variation of As concentration in European bottled water (and European water in general) covers four orders of magnitude (+2 orders of magnitude from natural As-provinces). As

21 As μg/l < 0.03 N 77 Probability % Relative frequency km Drinking water standard: 10 μg/l EU Directive 2003/40/EC Mineral Water μg/l μg/l NW NE SE SW μg/l Analytical Method: ICP QMS Detection limit: 0.03 μg/l 45 (5.09%) < detection limit n: 884 Minimum: < 0.03 μg/l 5%: < 0.03 μg/l 25%: μg/l Median: μg/l 75%: μg/l 95%: 5.15 μg/l Maximum: 89.8 μg/l Reimann & Birke (Eds.) (2010): Geochemistry of European Bottled Water ISBN Arsenic As

22 Appendix A 221 Appendix A Table of Water Action Levels as defined by the EU and for the majority of Countries from which water samples are included in this study Notation European Union Directives * Suitable for low Na diet (A) The maximum limit for boron will be fi xed, where necessary, following an opinion of the European Food Safety Authority and on a proposal from the Commission by 1 January (B) Natural mineral waters with a fl uoride concentration exceeding 1.5 mg/l shall bear on the label the words contains more than 1.5 mg/l of fl uoride: not suitable for regular consumption by infants and children under 7 years of age N1: Where possible, without compromising disinfection, Member States should strive for a lower value. For the water referred to in Article 6(1)(a), (b) and (d), the value must be met, at the latest, 10 calendar years after the entry into force of the Directive. The parametric value for bromate from fi ve years after the entry into force of this Directive until 10 years after its entry into force is 25 μg/l. N2: The value applies to a sample of water intended for human consumption obtained by an adequate sampling method (1) at the tap and taken so as to be representative of a weekly average value ingested by consumers. Where appropriate the sampling and monitoring methods must be applied in a harmonized fashion to be drawn up in accordance with Article 7(4). Member States must take account of the occurrence of peak levels that may cause adverse effects on human health. N3: For water referred to in Article 6(1)(a), (b) and (d), the value must be met, at the latest, 15 calendar years after the entry into force of this Directive. The parametric value for lead from fi ve years after the entry into force of this Directive until 15 years after its entry into force is 25 μg/l. Member States must ensure that all appropriate measures are taken to reduce the concentration of lead in water intended for human consumption asample much as possible during the period needed to achieve compliance with the parametric value. When implementing the measures to achieve compliance with that value Member States must progressively give priority where lead concentrations in water intended for human consumption are highest. N4: Member States must ensure that the condition that [nitrate]/50 + [nitrite]/3 # 1, the square brackets signifying the concentrations in mg/l for nitrate (NO 3 ) and nitrite (NO 2 ), is complied with and that the value of 0.10 mg/l for nitrites is complied with ex water treatment works. F.A.O. (1) If the product contains more than 1 mg/l of fluoride, the following term shall appear on the label as part of, or in close proximity to, the name of the product or in an otherwise prominent 5 Codex Standard position: contains fluoride. In addition, the following sentence should be included on the label: The product is not suitable for infants and children under the age of seven years where the product contains more than 1.5 mg/l fl uorides. W.H.O. * Short-term exposure ** Long-term exposure C = concentrations of the substance at or below the health-based guideline value may affect the appearance, taste or odour of the water, resulting in consumer complaints L = Large water treatment facilities N = No health-based guideline value for Cl in drinking water; >250 mg/l is detectable by taste O = ph value Optimum required range P = provisional guideline value, as there is evidence of a hazard, but the available information on health effects is limited S = Small water treatment facilities T = provisional guideline value because calculated guideline value is below the level that can be achieved through practical treatment methods, source protection, etc. pages