European Geographic Information Infrastructures: Opportunities and Pitfalls - GISDATA 5 (Gisdata, 5)

European Geographic Information Infrastructures

Also in the GISDATA Series I

GIS and generalization: methodology and practice edited by J.C.Müller, J.-P.Lagrange and R.Weibel

II

Geographic Objects with Indeterminate Boundaries edited by P.A.Burrough and A.U.Frank

III GIS diffusion: the adoption and use of geographical information systems in local government in Europe edited by I.Masser, H.Campbell and M.Craglia IV Spatial Analytical Perspectives on GIS edited by M.Fischer, H.J.Scholten and D.Unwin

Series Editors I.Masser and F.Salgé

European Geographic Information Infrastructures Opportunities and Pitfalls

EDITORS

PETER BURROUGH AND IAN MASSER

GISDATA V

SERI ES EDITORS

I.MASS E R and F.SALGÉ

UK

Taylor & Francis Ltd, 1 Gunpowder Square, London, EC4A 3DE

USA

Taylor & Francis Inc., 1900 Frost Road, Suite 101, Bristol, PA 19007 Copyright © Taylor & Francis Ltd 1998 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, electrostatic, magnetic tape, mechanical, photocopying, recording or otherwise, without the prior permission of the copyright owner. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. ISBN 0-203-21277-0 Master e-book ISBN

ISBN 0-203-26996-9 (Adobe eReader Format) ISBN 0-7484-0755-3 (hard) ISBN 0-7484-0756-1 (paperback) Library of Congress Cataloging-in-Publication Data are available Cover design by Hybert Design and Type, Waltham St Lawrence, Berkshire.

Contents

Series Editors’ Preface Editors’ Preface Contributors European Science Foundation 1

2

3

4

5

6

International aspects of spatial data exchange Peter Burrough and Ian Masser

vii ix xi xiii 1

PART ONE The Creation of Multinational Databases

15

From an understanding of European GI economic activity to the reality of a European data set François Salgé

17

Transboundary European GIS databases: review of the Baltic region experiences Sindre Langaas

31

PART TWO Multinational Databases for Specific Applications

45

Development of a GIS for hydrological modelling of the River Rhine Jaap Kwadijk and Erik Sprokkereef

47

The MEDALUS georeferenced database: an application used for land degradation research in the european mediterranean Erik Cammeraat and Hein Prinsen

57

Developments in cross-border standards for geodemographic segmentation Richard Webber

71 v

vi

CONTENTS

Multinational Database Products

85

European geoinformation data publishing: understanding the commercial industry Vanessa Lawrence

87

PART THREE 7

8

Data integration for commercial information products: experiences from the EC’s IMPACT-2 programme Roger Longhorn PART FOUR Legal and Institutional Issues Associated with the Development of Multinational Databases

9

Legal protection of geographic information in the EU Mireille M.M.van Eechoud

10 Intellectual property and mapping: a European perspective Laila Aslesen 11 Legal and institutional issues to be resolved with respect to the integration of European Data Andrew Larner 12 Intellectual property rights in disseminating digital geographic data, products and services: conflicts and commonalities among EU and US approaches Harlan Onsrud and Xavier Lopez Index

101

113 115

127

137

153

169

The GIS Data Series Editors’ Preface

Over the past few years there have been many signs that a European GIS community is coming into existence. This is particularly evident in the launch of the first of the European GIS (EG IS) conferences in Amsterdam in April 1990, the publication of the first issue of a GIS journal devoted to European issues (GIS Europe) in February 1992, the creation of a multipurpose European ground-related information network (MEGRIN) in June 1993, and the establishment of a European organisation for geographic information (EUROGI) in October 1993. Set in the context of increasing pressures towards greater European integration, these developments can be seen as a clear indication of the need to exploit the potential of a technology that can transcend national boundaries to deal with a wide range of social and environmental problems that are also increasingly seen as transcending the national boundaries within Europe. The GISDATA scientific programme is very much part of such developments. Its origins go back to January 1991, when the European Science Foundation funded a small workshop at Davos in Switzerland to explore the need for a European-level GIS research programme. Given the tendencies noted above, it is not surprising that participants of this workshop felt very strongly that a programme of this kind was urgently needed to overcome the fragmentation of existing research efforts within Europe. They also argued that such a programme should concentrate on fundamental research and it should have a strong technology transfer component to facilitate the exchange of ideas and experience at a crucial stage in the development of an important new research field. Following this meeting a small coordinating group was set up to prepare more detailed proposals for a GIS scientific programme during 1992. A central element of these proposals was a research agenda of priority issues grouped together under the headings of geographic databases, geographic data integration, and social and environmental applications. The GISDATA scientific programme was launched in January 1993. It is a fiveyear scientific programme of the Standing Committee of Social Sciences of the European Science Foundation. By the end of the programme more than 300 scientists from 20 European countries will have directly participated in GISDATA activities and many others will have utilised the networks built up as a result of them. Its objectives are: vii

viii

SERIES EDITORS′ PREFACE

¡ to enhance existing national research efforts and promote collaborative ventures over¡ ¡ ¡ ¡ ¡

coming European-wide limitations in geographic data integration, database design and social and environmental applications; to increase awareness of the political, cultural, organisational, technical and informational barriers to the increased utilisation and interoperability of GIS in Europe; to promote the ethical use of integrated information systems, including GIS, which handle socio-economic data by respecting the legal restrictions on data privacy at the national and European levels; to facilitate the development of appropriate methodologies for GIS research at the European level; to produce output of high scientific value; to build up a European network of researchers with particular emphasis on young researchers in the GIS field.

A key feature of the GISDATA programme is the series of specialist meetings that has been organised to discuss each of the issues outlined in the research agenda. The organisation of each of these meetings is in the hands of a small taskforce of leading European experts in the field. The aims of these meetings are to stimulate research networking at the European level on the issues involved and to produce high quality output in the form of books, special issues of major journals and other materials. With these considerations in mind, and in collaboration with Taylor & Francis, the GISDATA series has been established to provide a showcase for this work. It will present the products of selected specialist meetings in the form of edited volumes of specially commissioned studies. The basic objective of the GISDATA series is to make the findings of these meetings accessible to as wide an audience as possible to facilitate the development of the GIS field as a whole. For these reasons the work described in the series is likely to be of considerable importance in the context of the growing European GIS community. However, given that GIS is essentially a global technology most of the issues discussed in these volumes have their counterparts in research in other parts of the world. In fact there is already a strong UK dimension to the GISDATA programme as a result of the collaborative links that have been established with the National Center for Geographic Information and Analysis through the United States National Science Foundation. As a result it is felt that the subject matter contained in these volumes will make a significant contribution to global debates on geographic information systems research. Ian Masser François Salgé

Editors’ Preface

The origins of this book lie in the conversation that took place between the editors during a long car journey from Luxembourg to Wageningen in the Netherlands across several of Europe’s many border regions after a meeting organised by DGXIII of the European Commission in September 1995 to discuss one of the many drafts of the GI2000 document. During our discussion it emerged that, although it is generally accepted that there is a growing demand for multinational and pan-European databases, there is little hard information available on the problems encountered in different types of multinational geographic information applications, and there has been only limited discussion of the broader legal and institutional issues involved at the supranational level. With this in mind it was decided that a specialist meeting should be organised to explore these issues under the auspices of the European Science Foundation GISDATA scientific programme. It was envisaged that the products of such a meeting would make not only an important contribution to developing the European geographic information policy research agenda but also a much needed input to the ongoing GI2000 debates. The specialist meeting took place at Buoux in southern France in May 1996. It was organised by a taskforce consisting of Peter Burrough (University of Utrecht), Andrew Larner (then at the University of East London), lan Masser (University of Sheffield) and François Salgé (Institute Géographique National). The meeting was attended by 17 participants from an unusually wide range of backgrounds. These included not only the environmental and social sciences, surveying and law, but also representatives from regional government bodies, the EU itself, and international agencies such as MEGRIN and UN GRID. In contrast to most of the other specialist meetings organised under the auspices of the GISDATA programme, the private sector was strongly represented at Buoux. The presentations at the meeting fell into two categories: case studies of a variety of different applications involving geographic information through at least two countries, and broader perspectives on the legal and institutional issues that need to be considered at the European level. One of the most distinctive features of the meeting, particularly given the diversity of the backgrounds of the participants, was the degree to which there is general agreement on the nature of the issues that need to be considered. In this respect the findings of the case studies and the discussion of the broader legal and institutional issues complemented one another, the latter putting the issues raised in the case studies ix

x

EDITORS′ PREFACE

into a broader perspective, while the former provided specific examples of the consequences of these perspectives on current practice. Given the extent to which there was common ground between the participants, it was agreed that there should be two outcomes of this meeting. The first of these is this volume which presents the findings of the case studies and discusses the broader legal and institutional issues that need to be considered at the transnational level. The second was the preparation of a draft position statement ‘Geographic Information: The European Dimension’, which was placed on the World Wide Web and also circulated to key officials and decision-makers at the European level. This position statement identifies four activity areas in the field of geographic information: the mass market, operational functions, strategic functions and research. It considers the requirements of each of these from the standpoint of the key policy issues involved, the research dimension and the role of the European Union. On the basis of this analysis a number of proposals are put forward on behalf of the GISDATA programme for action with respect to the development of metadata services, the creation of core data sets for the EU as a whole, the need for common rules governing access to these and other data sets, and the contribution that can be made by the research community. A revised version of this position statement was subsequently adopted by the steering committee for the GISDATA programme and published both on the World Wide Web (htpp://www.shef.ac.uk/uni/academic/D-H/gis/policy.html) and in the GISDATA Newsletter (Burrough et al., 1997). In conclusion, the editors of this volume would like to express their gratitude to all those who contributed to the Buoux meeting and made it such a success. We would also like to thank Max Craglia for his helpful suggestions on the manuscript and Christine Goacher who took on the demanding task of typing and formatting the manuscript for publication. Peter Burrough and Ian Masser

REFERENCE BURROUGH, P., CRAGLIA, M., MASSER, I. and SALGÉ, F. (1997). Geographic information: the European dimension, GISDATA Newsletter, 8, 6–16.

Contributors

Laila Aslesen Statens Kartverk, 3500 Honefoss, Norway Peter Burrough Geographical Institute, University of Utrecht, PO Box 80.115, Heidelberglaan 2, 3508 TC Utrecht, The Netherlands Erik Cammeraat ICG, Landscape and Environmental Research Group, University of Amsterdam, Nieuwe Prinsengracht 130, NL-1018 VZ Amsterdam, The Netherlands Mireille M.M.van Eechoud Institute for Information Law, University of Amsterdam, Rokin 84, NL-1012 KX Amsterdam, The Netherlands Jaap Kwadijk Center for Geo-ecological Research, Department of Physical Geography, Utrecht University, PO Box 80.115, 3508 TC Utrecht, The Netherlands Sindre Langaas UNEP/GRID-Arendal do Dept of Systems Ecology, Stockholm University, S-106 91 Stockholm, Sweden Andrew Larner Local Government Management Board, Layden House, 76–86 Turnmill Street, London EC1M 5QU, UK Vanessa Lawrence Autodesk Ltd, Cross Lanes, Guildford GU1 1UJ, UK Roger Longhorn IDG Limited, EC Project Office, Neihaff, L-9161 Ingeldorf, Luxembourg xi

xii

CONTRIBUTORS

Xavier Lopez School of Information Systems and Management (SIMS), University of California, Berkeley, CA 94720–4600, USA Ian Masser Department of Town and Regional Planning, University of Sheffield, Western Bank, Sheffield S10 2TN UK Harlan Onsrud National Center for Geographic Information and Analysis (NCGIA), University of Maine, Orono, Maine 04469–5711, USA Hein Prinsen ICG, Landscape and Environmental Research Group, University of Amsterdam, Nieuwe Prinsengracht 130, NL-1018 VZ Amsterdam, The Netherlands François Salgé IGN, 136 bis Rue de Grenelle, 75007 Paris 07SP, France Erik Sprokkereef CHR/KHR-GIS Group, RIZA-Arnhem, PO Box 9072, 6800 ED Arnhem, The Netherlands Richard Webber Experian, Talbot House, Talbot Street, Nottingham NG1 5HF, UK

The European Science Foundation is an association of its 55 member research councils, academies, and institutions devoted to basic scientific research in 20 countries. The ESF assists its member organisations in two main ways: by bringing scientists together in its scientific programmes, networks and European research conferences, to work on topics of common concern, and through the joint study of issues of strategic importance in European science policy. The scientific work sponsored by ESF includes basic research in the natural and technical sciences, the medical and biosciences, and the humanities and social sciences. The ESF maintains close relations with other scientific institutions within and outside Europe. By its activities, the ESF adds value by cooperation and coordination across national frontiers and endeavours, offers expert scientific advice on strategic issues, and provides the European forum for fundamental science. This volume is the fifth in a series arising from the work of the ESF Scientific Programme on Geographic Information Systems: Data Integration and Database Design (GISDATA). This programme was launched in January 1993 and through its activities has stimulated a number of successful collaborations among GIS researchers across Europe. Further information on ESF activities in general can be obtained from: European Science Foundation 1 quai Lezay Marnesia 67080 Strasbourg Cedex tel: +33 88 76 71 00 fax: +33 88 37 05 32

xiii

EUROPEAN SCIENCE FOUNDATION This series arises from the work of the ESF Scientific Programme on Geographic Information Systems: Data Integration and Database Design (GISDATA). The Scientific Steering Committee of GISDATA includes: Dr Antonio Morais Arnaud Faculdade de Ciencas e Tecnologia Universidade Nova de Lisboa Quinta da Torre, P-2825 Monte de Caparica Portugal

Professor Michael F. Goodchild National Center for Geographic Information and Analysis (NCGIA) University of California Santa Barbara, California 93106 USA

Professor Hans Peter Bähr Universität Karlsruhe (TH) Institut für Photogrammetrie und Fernerkundung Englerstrasse 7, Postfach 69 80 (W) 7500 Karlsruhe 1 Germany

Professor Einar Holm Geographical Institution University of Umeå S-901 87 Umeå Sweden

Professor Kurt Brassel Department of Geography University of Zurich Winterthurerstrasse 190 8057 Zurich Switzerland

Dr Massimo Craglia (Research Coordinator) Department of Town & Regional Planning University of Sheffield Western Bank, Sheffield S10 2TN UK

Professor Jean-Paul Donnay Université de Liège, Labo. Surfaces 7 place du XX août (B.A1–12) 4000 Liège Belgium

Professor Manfred Fischer Department of Economic and Social Geography Vienna University of Economic and Business Administration Augasse 2–6, A-1090 Vienna, Austria xiv

Professor lan Masser (Co-Director and Chairman) Department of Town & Regional Planning University of Sheffield Western Bank, Sheffield S10 2TN UK Dr Paolo Mogorovich CNUCE/CNR Via S.Maria 36 50126 Pisa Italy Professor Nicos Polydorides National Documentation Centre, NHRF 48 Vassileos Constantinou Ave. Athens 116 35 Greece M.François Salgé (Co-Director) IGN 2 ave. Pasteur, BP 68 94160 Staint Mandé France Professor Henk J.Scholten Department of Regional Economics Free University De Boelelaan 1105 1081 HV Amsterdam Netherlands

EUROPEAN SCIENCE FOUNDATION

Dr John Smith European Science Foundation 1 quai Lezay Marnesia F67080 Strasbourg France Professor Esben Munk Sorensen Department of Development and Planning Aalborg University, Fibigerstraede 11 9220 Aalborg Denmark

Dr Geir-Harald Strand Norwegian Institute of Land Inventory Box 115, N-1430Ås Norway Dr Antonio Susanna ENEA DISP-ARA Via Vitaliano Brancati 48 00144 Roma Italy

xv

CHAPTER ONE

International aspects of spatial data exchange PETER BURROUGH AND IAN MASSER

Context The increasing globalisation of national economies and the need for concerted international efforts in the environmental field are creating a demand for transnational spatial data. Three key issues—technology, disciplinary aspects, and organisation—control the ways that spatial data from different sources can be successfully brought together for a common good. The first issue, technology, addresses the fact that there are many different ways of storing and transmitting spatial data in electronic form. To meet demands for improving the technical aspects of data exchange there have been calls for the establishment of transnational and even global spatial data infrastructures and spatial data standards to improve system interoperability (see, for example, McKee, 1996; Schell, 1995). The second issue, disciplinary aspects, addresses the problems arising from the fact that related sciences and technical methods do not necessarily describe or recognise the same spatial phenomena in the same way; these problems are being addressed by the development of international standards for data description. With respect to transnational spatial data, the essential issue is the recognition that most spatial data are collected by national rather than supranational agencies thereby reflecting local and national rather than wider regional or global concerns. For all these reasons a great deal of non-technical effort is often required to obtain and to transform national data sets on similar or related themes into spatial databases suitable for international applications. In the process, a wide range of legal and institutional issues may have to be resolved in addition to the technical and disciplinary problems involved. This growing demand for transnational data sets is particularly evident in the European Union. Over the past decade the EU has been moving towards a single European market to promote the free flow of goods, services, people and capital among its fifteen member states. The EU has also been very active on the environmental front and its CORINE (Coordinated Information on the European Environment) programme represents a major pioneering effort to establish a multi-source, multinational environmental GIS (Mounsey, 1991). Since 1990 there has been a growing European dimension to spatial planning with the publication of the Europe 2000 report on the development of the Community’s territory (CEC, 1991) and the Europe 2000+report on cooperation for European trerritorial development (CEC, 1994). 1

2

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Two levels of demand for transnational data can be identified with respect to the European experience. At the most basic level there is a demand for data to assist in the provision of basic services and the coordination of planning efforts in cross-border areas. Given the number of member states involved and the specific geographical circumstances of Europe, a large proportion of the European land mass falls into this category. The Europe 2000 report, for example, estimates that ‘the regions along these borders account for around 15% of the total land area of the Community and some 10% of its population’ (CEC, 1991, p. 166). These regions vary widely in terms of language, economic, cultural and physical characteristics. For example, the frontiers of Spain and France and France and Italy are largely mountainous areas with relatively small populations whereas those among the Dutch/ German, Dutch/Belgium and Belgium/French borders are not divided by any dominant physical features and have relatively large populations. A number of cross-border data integration projects are already under way in these border regions. For example, an integrated geographic database has already been created for the Pyrenees area by the GIS subcommission of the Pyrenees Work Community which includes Andorra, three French regions and four Spanish regions (Morant, 1996). Similarly, the feasibility of a coordinated GIS for administrative purposes is being explored by the Dutch provinces of Groningen and Drenthe together with their counterparts on the German side of the border (GAG, 1995), and an integrated GIS has been created for forest management on the German, Austrian and Czech borders by the Bavarian Ministry for Regional Environmental Affairs, the Czech Ministry of Economic Affairs and the Office of the Upper Austrian Provincial Government (Schaller et al., 1995). The other level of demand for transnational data falls somewhere between cross-border initiatives such as these and pan-European projects such as CORINE. Typically, this level of demand involves a large number of countries and/or a larger scale of operations than that involved in the efforts described above. Historically, environmental issues have dominated applications at this level but recently there has been an increase in strategic planning initiatives. From Figure 1.1 it can be seen that many of the environmental projects involve river basins such as the Rhine (see Chapter 4 of this book), or maritime regions such as the Baltic (see Chapter 3), the North Sea or the Mediterranean. Some of the planning projects at this level also reflect such divisions. This is the case, for example, in the database that is being developed for the Vision and Strategies around the Baltic Sea 2010 project by the Council of Baltic Sea States (see also Chapter 3 of this book). However, other projects reflect traditional national boundaries, as can be seen in the work that has been undertaken in connection with the preparation of a structural plan for the Benelux countries as a whole (Zonneveld and D’hondt, 1994). Although the examples listed above refer predominantly to public sector projects, similar developments are under way in the private sector particularly with respect to the creation of a European digital road network database and in the field of geodemographics (Bastiaansen, 1995). In connection with the former there are several projects under way to develop commercial digital road network databases for Europe as a whole (Ireland, 1995). The demand for the latter has increased considerably in the past few years as the number of multinational marketing operations has expanded (Birkin, 1996; see also Chapter 6 of this book). The overall demand for other European-level products has also grown alongside these specific developments. As a result a number of major publishers have entered the European market (see Chapter 7). For example, Bartholomew launched its first commercial comprehensive digital atlas for Europe at a scale of 1:1,000,000 on CD ROM in early 1996 (Bartholomew, 1996) and the EU itself has taken steps to promote the European geographic

INTERNATIONAL ASPECTS OF SPATIAL DATA EXCHANGE

3

Figure 1.1 Important transborder regions in Europe.

information market through its Information Market Policy Actions programme (IMPACT) (see Chapter 8 of this book). Despite these developments, however, only limited progress has been made towards the coordination of geographic information at the European level and the creation of supranational data sets. Given the growing demand for data sets of this kind it can be argued that this represents a major barrier to management and policy-making in situations where more than one country is involved. Some of the key features of these arguments are summarised in the report The Social Sciences in the Context of the European Communities which was commissioned by DGXII (Research and Development) (ESF/ESRC, 1991). This concluded that: the benefits of social science to the Community are constrained by the lack of social science data in an appropriate form. On the national level, Europe is data rich. In most countries there is a well established system of market and opinion research institutes, statistical offices, academic social research institutes and, to a varying extent, the social science data service infrastructure. In addition public administration contributes to its database an enormous amount of administrative data. In spite of this data wealth, research with European perspectives is seriously handicapped: the European database is not well integrated, large scale research is hardly coordinated, measurement instruments and data representation lack compatibility, data access and data protection regulations differ, and even information about the availability of data is not always easy to obtain, (p. 74)

4

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 1.1 Examples of variations in the estimation of land cover in Europe (‘000 km2)

Source: RIVM (1994)

This report was basically concerned with social science research but there are clear parallels to these arguments in the field of environmental research. Table 1.1, for example, illustrates some of the differences in interpretation of land cover at the European level with reference to data from various pan-European sources. From this it can be seen that the magnitudes of the differences between the various estimates greatly exceed the EU’s own norms for the accuracy of recording land use which are used for computing agricultural support payments through the Integrated Agricultural Census System (IACS) collected under European law (Aspinall, 1997). The demand for more compatible data on the different European countries has prompted a number of initiatives at the European level by ad hoc bodies and also by the EU itself. Of particular importance in the case of the former is the tradition of informal cooperation between the national mapping agencies that has been built up under the auspices of CERCO (Comité Européen des Responsables de la Cartographie Officielle). An important outcome of this cooperation is the initiative to establish a multipurpose European ground-related information network (MEGRIN; see Chapter 2) and the continuing efforts of the two technical GIS subcommittees of the European Standards Commission (CEN) to promote common standards in the field of geographic information. There have also been a number of significant initiatives within the European Commission (EC) itself in the past few years to make data more European. The European Statistical Office (EUROSTAT) has played a key role in the harmonisation of statistics collected by the national statistical agencies of the EU member states and the Directorate General for Telecommunications, Information Markets and the Exploitation of Research (DGXI II) has been particularly active in the field of geographic information. This Directorate was instrumental in the creation of the European Umbrella Organisation for Geographic Information (EUROGI) in 1993. The aim of EUROGI is ‘not to replace existing organisations but to catalyse effective cooperation between existing national, international and discipline oriented bodies [such as CERCO] to bring added value in the areas of Strategy, Coordination, and Services’ (Burrough et al., 1993, p. 31). The most important contribution of DGXIII and the Commission as a whole to date has been in the preparation of the GI2000 document (CEC, 1996). Discussion of this document

INTERNATIONAL ASPECTS OF SPATIAL DATA EXCHANGE

5

within the Commission is intended to secure a commitment from the Council of Ministers for an integrated programme of actions designed to put in place a European policy framework for geographic information. Prior to its entry into the Commission’s machinery, the GI2000 document was the subject of numerous rounds of consultations with important stakeholders in geographic information throughout Europe. In the process, the original document went through at least seven drafts and its title changed from ‘GI2000: Towards a European Geographic Information Infrastructure’ to ‘GI2000: Towards a European Policy Framework for Geographic Information’ (Craglia and Masser 1997). Given its importance in the current debate, it is useful to summarise the main arguments underlying this document. The GI2000 document presents a policy framework for European geographic information which highlights the need for political action to overcome the barriers identified above: The major impediments to the widespread and successful use of geographic information in Europe are not technical, but political and organisational. The lack of a European mandate on geographic information is retarding development of joint information strategies and causes unnecessary costs, is stifling new goods and services and reducing competitiveness. (p. 1) What is required, then, is ‘a stable, European-wide set of agreed rules, standards, procedures, guidelines and incentives for creating, collecting, exchanging and using geographic information’ (p. 2). The GI2000 document points out that many elements of a European geographic information infrastructure already exist. There is a growing number of high-quality digital geographic databases in Europe held by local, national and European providers and users in both the public and private sectors. However, to convert these assets into an effective European geographic information infrastructure requires the following political decisions:

¡ Agreement of member states to set up a common approach to create European base data, and to make this generally available at affordable rates.

¡ A joint decision to set up and adopt general data creation and exchange standards and to use them.

¡ A joint decision to improve the ways and means for both public and private agencies and similar organisations to conduct European level actions, such as creation of seamless pan-European data sets. ¡ Agreement and actions to ensure that European solutions are globally compatible, (p. 2) The GI2000 document also lists a number of practical actions that will be needed to bring a European geographic information structure into being. These include: ¡ stimulating the creation of base data which is regarded as ‘the single most important area to the development of the market for geographic information’ (p. 15); ¡ stimulating the creation of metadata services to overcome the general lack of awareness of what data exist; ¡ lowering legal barriers and reducing potential risks; ¡ stimulating public/private synergy; and ¡ coordination between member state policies at EU level and between the actors in the marketplace. As might be expected, there are many parallels between the issues raised in the GI2000 document and those raised by national agencies such as the United States Federal Geographic Data Committee which has been charged by President Clinton with the establishment of the National Spatial Data Infrastructure (Executive Office of the President, 1994) and the Dutch

6

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

National Council for Geographic Information which has produced a discussion paper on its national geographic information infrastructure (RAVI, 1995). What distinguishes the GI2000 document from these other initiatives, however, is that it deals with the transnational rather than the national dimension of geographic information strategy.

The purpose of this book One feature of much of the current debate on the European geographic information infrastructure is the lack of systematic information on the nature of the problems experienced by different types of GI user and/or supplier in developing databases or products that involve data from more than one country within Europe. Another feature is the relatively limited discussion that has taken place on the legal and institutional issues involved at the European as against the (intra)national level. With these considerations in mind, this book attempts to fill these gaps. It presents the findings of a number of case studies of multinational geographic information projects within Europe and puts forward a number of different perspectives on the legal and institutional issues that need to be resolved before a Europe-wide infrastructure can be effective. In the process it makes a contribution to the debates that are taking place at the European level and also draws attention to some of the policy research issues associated with them. It is worth noting at this stage that the subtitle of the book refers to the opportunities and pitfalls associated with the development of a European geographic information infrastructure. There are two main forces creating new opportunities for geographic information at present. The first of these is the growing demand for cross-border and multinational data that is associated with the increasing globalisation of economic activity in general and the establishment of supranational bodies, such as the European Union in particular. The second is the increasing availability of cheap, powerful information and communications technology which has already transformed conventional mapping and cartography out of all recognition. In place of the traditional map this new technology makes it possible to integrate geographic information from a variety of different sources to specifications designed by users rather than producers. Consequently, as the GI2000 document points out, ‘each new GI data set could be a plug-in module to a wider European collection of data sets. The critical elements in this idea are no longer computer hardware and software but the availability, price, quality and technical structure of the data, and legal aspects covering its dissemination and re-use’ (CEC, 1996, p. 9). For this reason most of the discussion in the book revolves around the factors governing the dissemination and re-use of data collected by other agencies. These represent the ‘pitfalls’ or the issues that must be resolved before an effective European geographic information infrastructure can come into operation. In this respect the contributors to the book share the views expressed in the GI2000 document that ‘the major impediments of the widespread and successful use of geographic information in Europe are not technical, but political and organisational’ (p. 1). As a result, the discussion is in essence complementary to the ongoing technical debate led by organisations such as the Open GIS™ Consortium (OGIS) on the need for greater interoperability (Schell, 1995; Glover, 1996). It should be also noted that, although this book deals primarily with multinational geographic information initiatives from a European perspective, many of the lessons that can be learnt from it are of relevance to any situation where applications are envisaged which involve integrating data derived from more than one country. For this reason, then, much of this experience is relevant to global as well as European circumstances.

INTERNATIONAL ASPECTS OF SPATIAL DATA EXCHANGE

7

Overview The remainder of this book contains eleven chapters written by authors from a wide range of backgrounds who were specially selected for their experience in multinational European spatial data issues. The first seven of these present the findings of a series of case studies of multinational geographic information projects in Europe while the last four present different perspectives on the legal and institutional issues involved.

Findings of the case studies For purposes of presentation the case studies are grouped into three categories describing the main types of initiative involved: the creation of multinational databases for multipurpose use, the development of multinational databases for specific applications, and the development of multinational database products. A separate section evaluating the findings as a whole follows the overview of the case studies themselves.

The creation of multinational databases Two contributions consider the issues involved in the development of multipurpose multinational databases. In the first of these, François Salgé describes the work of MEGRIN (Multipurpose European Ground Related Information Network), whose partners are all official mapping agencies. MEGRIN was set up primarily to stimulate the market for European geographic information and to facilitate the distribution of geographic information produced by the partners in the various European countries. To promote better awareness of the availability of European digital information projects, MEGRIN launched its Geographical Data Description Directory (GDDD) metadata service in 1993. Since then it has taken the first step towards creating a Europe-wide data set by the development of the Seamless Administrative Boundaries of Europe (SABE) data set. SABE is a vector data set which has been created from administrative boundary data provided by the national mapping agencies of over twenty European countries including Iceland, Cyprus, Switzerland and Latvia as well as the fifteen members of the EU. To create such a data set a number of technical problems have to be resolved. These include the transfer of existing data into a unique geodetic referencing system and the creation of a unique geometric and topologic frame. There are also a number of quite complex conceptual issues to be tackled given that each of the participating countries has its own distinctive administrative hierarchy. In this case the question of the need for a template frame was limited to the datum of 1 January 1991, thereby avoiding a number of questions relating to the update of the database. As a result of these efforts, two versions of SABE 91 are currently available: SABE 30 which corresponds to an approximate scale of 1:100,000 and a derived product, SABE 200 which corresponds to an approximate scale of 1:1,000,000. The former costs ECU 130,000 and the latter ECU 15,000. The key principle that underlies these prices is that the overall price of the data set is governed by the combined prices of the equivalent national data sets. There are many similarities between the experience of SABE and that relating to the development of seamless multilayer databases for the Baltic region which is described by Sindre Langaas in Chapter 3. The experience of the Baltic Region is particularly interesting in this respect because there is a long tradition of international cooperation in the

8

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

environmental field since the signing of the Helsinki Convention in 1974. More recently there has also been a growing interest in the region as a whole as a result of an increase in multinational spatial planning activities and the demands from multinational companies for market-relevant data for the whole region. Langaas reviews the experience of two small-scale, coarse resolution projects to establish thematic databases for the Baltic region. The first of these utilises information from about twenty of the data sets in the Global Resource Information Database (GRID) which is held in GRID-GENEVA to create a package for GIS education, research and planning. The second of these projects, the Baltic Drainage Basin Project, is a more ambitious programme involving fourteen countries including many members of the former Soviet Union as well as some members of the EU. The basic objective of this collaborative research project was to create a seamless, multi-thematic and coherent GIS database for the whole of the Baltic region. The findings of Langaas’ evaluation show that, although most of the data sets required were made available free of charge or were already in the possession of the developers, it was still necessary in a number of other cases to take steps to resolve copyright issues relating to them. This reflects the lack of practical rules of conduct and legal practices for the redistribution of multinational GIS data sets derived from small-scale and coarse resolution data originating from multiple sources. The findings of the evaluation show that the costs involved in such a task are not trivial even though no primary data collection was involved. In practice it was found that the costs of data conversion to a single data model were often substantial and that these were compounded in many cases by the lack of metadata describing the data sets themselves. The success of these two projects, in the opinion of Langaas, is largely owing to the fact that both of them had a clear environmental focus. The experience of these and other projects in the Baltic region suggest that focused efforts driven by one or more applications are more likely to succeed than multipurpose efforts. The findings of the evaluation also suggest that the size of the group involved may be an important factor, and that the larger the group, the more likely it is that the range of applications will increase.

Multinational databases for specif ic applications Three case studies review the experiences of creating multinational data sets for specific applications. In general the findings of these confirm Langaas’ conclusion that focused efforts which are driven by specific applications are likely to encounter fewer problems than multipurpose initiatives involving a large number of interests. In the first case study in Part Two, Jaap Kwadijk and Erik Sprokkereef describe the development of RHINEFLOW, a distributed runoff model specially written to study the reaction of the discharge pattern of the River Rhine to variations in precipitation and temperature (the latter from global warming). Modelling hydrological and environmental processes on a scale such as the Rhine basin requires the integration of a great deal of data from many sources. From the outset of their project the authors recognised that it was unlikely that all the data required would be available and that it would be necessary to make use of simplified descriptions of complex economic and social processes. The first version of the RHINEFLOW model was developed as part of a doctoral dissertation. It drew upon a limited range of accessible data resources. A second, extended version of the model at a higher spatial and temporal resolution is now under way under the auspices of the Commission for the Hydrology of the Rhine (CHR), the research arm of the International Rhine Commission. It is intended that this will be an operational

INTERNATIONAL ASPECTS OF SPATIAL DATA EXCHANGE

9

version which can be used over a long time by the different organisations involved. The CHR functions as an intermediary between the researchers and the data providers in this project, given that its members are representatives of national institutes for water management. Despite this, it is recognised that the inclusion of new variables in the model together with a higher level of resolution is likely to require a lot more effort on the part of the research team. It will be necessary, for example, to reconcile data collected from maps that have various scales, use different reference levels and originate from databases which use different classifications. Given the extent to which their data needs have changed over time, Kwadijk and Sprokkereef feel that it is not realistic to expect that any overarching database could be built which would include all the information needed to meet their present and future needs. They are also rather pessimistic about the extent to which there will be agreement on the classification and description of environmental data in the near future. For this reason they argue that the most effective step forward that can be made at the national and European levels is in the development of metadata facilities to help researchers like themselves find the data that meet their needs. Cammeraat and Prinsen’s contribution (Chapter 5) describes some features of the MEDALUS georeferenced database that has been developed as part of a large-scale collaborative multinational research project on land degradation in the European Mediterranean. Unlike the projects described in the previous two chapters, the eight core field sites involved in this case are not contiguous and include locations in France, Greece, Italy, Spain and Portugal. This project also differs from those described earlier in that most of the data were collected by the research teams themselves. The main problem in this case was to develop protocols for the standardisation and harmonisation of data collection and processing. Nevertheless, the teams also made extensive use of topographic and geological data and were compelled to obtain official permission to scan topographic maps and aerial photographs. This led to problems as some countries appeared to have no official policy, whereas others, like the IGN in France, had well developed commercial policies in this respect. A number of useful lessons can be learnt from this experience. The problems of transnational data integration are considerably simplified when they are internalised within project teams who collect a great deal of the data themselves to an agreed model specification. However, even in this case the group encountered problems with respect to the acquisition of data products from different data producers. Although these problems were resolved in most cases, the amount of time and effort involved in making the necessary arrangements was considerable. At first sight there are few similarities between the subject matter of Chapter 6 by Richard Webber and those of previous chapters in Part Two. Unlike the authors of the two previous chapters who are based in an academic research environment, Webber comes from the world of business. His work is primarily concerned with the development of geodemographic tools to assist companies who wish to tailor the marketing of their products to different types of customer. Webber describes the changing business context of marketing of the past 30 years which has shifted from the mass marketing of products to a relatively undifferentiated audience towards micro-marketing strategies addressed to specific niche audiences in particular localities. Despite the differences in starting points, however, there are many similarities between the approaches used by geodemographic analysts and those of environmental scientists such as Kwadijk and Sprokkereef or Cammeraat and Prinsen. Both groups have a clear idea of what they hope to achieve: environmental scientists wish to model hydrological or land degradation processes, while geodemographic analysts want to undertake cluster analyses of small area statistical information derived typically from census sources

10

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

supplemented by other data. Like the environmental scientists’ models, the nature of the classifications they are developing largely dictate the choice of data sources and, if data from standard sources are not available, they find alternative sources themselves. Webber points out that there has been a growing demand for European-level geodemographic decision support systems over the past few years as the number of multinational marketing operations in Europe has increased. In the last section of his chapter he describes the EuroMOSAIC product consisting often generalised residential neighbourhood types that has been developed as a common international standard for this purpose. Despite these considerable achievements, however, ‘there is still significant scope for harmonisation of European data sources to facilitate greater integration and utilisation within geodemographic analysis’. Multinational database products Two case studies consider some of the issues involved in the development of multinational data products for a mass market. In the first, Vanessa Lawrence explores the nature of the emerging commercial data publishing market in Europe and considers its future. Data publishing differs from conventional text publishing in that there are potentially many different owners of the intellectual property rights of both the software and data used in such products. Unlike the case with respect to text publishing where there is a well established model governing the intellectual property rights (IPR) of those involved, there is as yet no model for the electronic data publishing industry and policies differ considerably from country to country. Consequently, the key to publishing success is to own as much of the value chain as possible in order to minimise the risks involved in what is currently a very immature market where products are expensive to construct and the market for them is often unproven. The second case study in Part Three also considers the problems which face the developers of products involving geographic information for the mass market. In Chapter 8 Roger Longhorn evaluates the experience of a number of projects that were undertaken as part of the European Commission’s IMPACT-2 programme. The findings of his evaluation show that the problems facing the product developers fell mainly into two categories: market-related issues including legal and IPR issues, and technical questions. The former includes producing products for which the data cannot legally be sold as a result of restrictions imposed by the original data collectors, the need to take account of multiple data access policies across multiple data providers, and the lack of cooperation from data holders who see a product or a service more as a competitor than a potential source of revenue. In contrast, most of the technical data integration issues revolved around standards and the procedures used for attaching geographic references to attribute data. On the basis of his evaluation, Longhorn argues that future product developers should take note of this experience when planning their own GI-related products for a mass market. Like Lawrence, he argues that the key to success is the ownership of IPR: ‘if at all possible—own your own data sources! Or if that is not possible, then sort out ownership, sale and resale rights very early in the project’.

Evaluation Taken as a whole, the findings of the case studies highlight some of the problems that need to be resolved in the process of developing multinational databases for both general and

INTERNATIONAL ASPECTS OF SPATIAL DATA EXCHANGE

11

specific purposes and also multinational database products. These problems can usefully be grouped into four categories:

¡ data acquisition: problems encountered in acquiring geographic information from several ¡ ¡

¡

European countries and also in disseminating geographic information to users in other European countries (see, for example, Langaas); data usage: problems associated with variations in conditions of access between European countries and also problems encountered in creating multinational geographic information products as a result of variations in access (see, for example, Longhorn); data derivation: problems associated with the derivation of data sets for either internal use or external distribution from geographic information provided by several European countries and also those involved in deriving new data products from this information (see, for example, Lawrence or Salgé); data description: problems associated with locating data from other countries and obtaining enough information to decide whether it is appropriate for the envisaged task or whether there are potential risks such as liability for misrepresentation associated with its use (see, for example, Kwadijk and Sprokkereef).

At the same time, the findings of the case studies show that the effects of these problems were much less marked in cases where multinational databases were being developed for specific applications. In other words, all things being equal, focused efforts driven by specific applications are likely to encounter fewer problems than multipurpose initiatives involving a large number of interests or initiatives directed to the development of mass market products.

Legal and institutional issues associated with the development of multinational databases Questions relating to copyright, ownership, intellectual property rights, access and redistribution rights featured prominently throughout the case studies, and Part Four deals directly with the legal and institutional issues that face the developers of transnational and multinational products in Europe. In the first of the four chapters in Part Four, Mireille van Eechoud summarises some of the main findings of a comparative study of legal protection in the fifteen EU countries which was commissioned by EUROGI. Given that EUROGI often acts as the voice of the European geographic information industry in discussions with the EU, the study is significant in itself as it is the first legal and institutional issue to be addressed by that organisation. In essence, as Van Eechoud points out, a study of this kind is of considerable political importance because, ‘as a European market for geographic information develops, so does the need for adequate and harmonised legal protection of geographic data and information products’. With these considerations in mind she describes the present position relating to copyright and neighbouring rights law in the fifteen EU countries and then moves on to consider the extent to which the law on unfair competition can be used to protect the producers of geographic information and the degree to which specific measures have been enacted to protect geographic information producers in these countries. The findings of her analysis suggest that the problems experienced by the authors of earlier chapters only scratch the surface of the complex range of legal and institutional issues that need to be resolved in Europe as a whole before the potential of a fully operational geographic information

12

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

infrastructure can be exploited. As she points out, ‘for an adequate legal framework to exist, first and foremost transparency is required with regard to what type of geographic information is subject to (exclusive) rights, to what these rights pertain, and who is the beneficiary of the rights’. In the light of her survey of the different types of legal protection for geographic information in the EU member states, she concludes that at present there is ‘no such thing as a transparent system’. The contributions from Laila Aslesen and Andrew Larner also deal with matters relating to rights but from two contrasting positions. Aslesen draws upon her own experience of the Norwegian Mapping Agency in her examination of intellectual property rights and mapping (Chapter 10). She regards intellectual property protection as essential if agencies are to invest substantial sums of money in database creation. Geographic databases also differ from many other databases in that their value falls off dramatically after a year or two unless they are updated regularly. Under these circumstances the Norwegian Mapping Agency has developed an administrative and pricing policy which seeks to make data available without compromising data protection. This involves a differential scale for the sale of rights for usage for internal and/or private purposes as against the sale of marketing rights to developers of derived products based on their data. Like Van Eechoud, Larner (Chapter 11) argues that a clear system of rights in data must be defined before there can be any substantial degree of data integration at the European scale. However, he also feels that it is important for those involved with geographic information to identify what is needed before considering the current legal position. His analysis takes account of the needs of data users and data subjects as well as those of data producers such as national mapping agencies. Larner also argues not only that it is important to establish a system which attempts to find a balance between these rights but also that it will be necessary to create a system for enforcing these rights in a way that respects the interests of each of the parties involved. In Larner’s opinion, copyright law, even though it provides considerable protection to data providers in some European countries such as Britain and Ireland, is not appropriate for geographic data sets. It works best when creativity is involved which results in easily distinguishable products, whereas geographic information products must have a high degree of similarity to be used. Like Aslesen, Larner argues that the most appropriate way forward is for the development of some form of compulsory licensing scheme which can accommodate a multiplicity of rights without compromising the position of the data producer. In the last of the four chapters in Part Four, Harlan Onsrud and Xavier Lopez compare European and American experience with respect to legal protection and data availability. The main finding of their analysis is that ‘there appears to be widespread and growing belief in the USA that the general US approach to the treatment of copyright seems more appropriate in the long run for spurring technological innovation and economic vitality for an economy than the alternatives offered by the European Union to date’. The core of their argument rests on making a clear distinction between the public and private sectors with respect to geographic information. In so far as the private sector is concerned, they accept most of the previous arguments about the need for a system of intellectual property rights to provide an incentive for companies to create and distribute spatial data sets. They also note the limitations of current copyright law with respect to protection of geographic information and provide some examples of the increasing use of licensing arrangements to control the use and reproduction of spatial data sets by suppliers. However, their position with respect to the public sector is completely different. They point out that US public

INTERNATIONAL ASPECTS OF SPATIAL DATA EXCHANGE

13

information principles guarantee open access to public information as a precondition of both economic vitality and political accountability. Consequently, data collected by the Federal Government are exempt from copyright law and federal agencies are obliged to disseminate these data at only the costs of dissemination. It is a moot point whether such a hard and fast distinction between the public and private sectors is tenable in Europe where the differences between the two are becoming increasingly blurred as a result of the pressures on many government agencies to adopt market-orientated management styles and the long-standing tradition of using semi-public companies for a wide range of tasks including data infrastructure development. Nevertheless, the findings of Onsrud and Lopez’s analysis raise some interesting questions for further research at the European level. As might be expected, there are references in all four chapters (9–12) to the effectiveness of current European initiatives towards harmonisation and greater transparency with respect to geographic information, particularly in the context of the recent directive on the legal protection of databases. Whereas Onsrud and Lopez feel that such a directive over-protects database owners at the expense of users who wish to acquire and use data from (exclusive) suppliers, Van Eechoud and Larner take a less critical stance, welcoming the extent to which some measure of agreement has been reached on these matters at the European level.

Conclusion In summary, then, the demand for transnational data of all kinds is growing, particularly in the EU. However, most data are collected by national agencies for national purposes and many problems must be resolved before transnational databases can be created. These are primarily political and organisational rather than technical in nature and cover a wide range of legal and institutional issues. In practice, however, their impact varies according to the nature of the project and focused efforts driven by specific applications seem to encounter fewer problems than multipurpose initiatives involving a large number of different interests or the development of products for the mass market. The findings of the case studies described in this book suggest that these problems are exacerbated at the present time by a number of short-term factors. These include the absence of clear rules or agreed practices for the acquisition and re-use of data, the lack of adequate metadata to help users find the information they need, and the immaturity of the market for transnational geographic information products itself. The development of geographic information infrastructures also raises a number of more fundamental questions relating to the nature of the role of government with respect to geographic information. These include the nature of the legal instruments required to protect the rights of all those involved in the creation and maintenance of digital databases and the trade-off for governments between the need to make data freely available to promote technological innovation and ensure political accountability, and at the same time trying to recover some or all of the costs associated with digital database creation and maintenance. Problems such as these point to new agendas for geographic information research that are driven by political and organisational rather than technical considerations. In this respect the multinational initiatives described in this book represent a special case of these general agendas.

14

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

REFERENCES ASPINALL, R. (1997). Measurement of area in GIS: A rapid method for assessing accuracy of area measurement, in Kemp, Z. (Ed.), Innovations in GIS 4. London: Taylor and Francis. BARTHOLOMEW (1996). Euromaps on CD-ROM. London: Bartholomew Data Sales. BASTIAANSEN, A. (1995). Pan European geographic data, the critical success factor for GIS business applications, Proceedings of GIS for Business, pp. 95–7, Cambridge: Geoinformation International. BIRKIN, M. (1996). Selling in a single market, GIS Europe, 5(4), 20–2. BURROUGH, P., BRAND, M., SALGÉ, F. and SCHULLER, K. (1993). The EUROGI vision, GIS Europe, 2(4), 30–1. CEC (1991). Europe 2000: Outlook for the Development of the Community’s Territory. Brussels: DGXVI. CEC (1994). Europe 2000+: Cooperation for European Territorial Development. Brussels: DGXVI. CEC (1996). GI2000: Towards a European Policy Framework for Geographic Information. Luxembourg: DGXIII. CRAGLIA, M. and MASSER, I. (1997). Building the European geographic information resource base: towards a policy driven research agenda, in CRAGLIA, M. and ONSRUD, H. (Eds), Geographic Information Research: Transatlantic Perspectives. London: Taylor and Francis. ESF/ESRC (1991). The Social Sciences in the Context of the European Communities: A Report Commissioned by DGXII. Strasbourg: European Science Foundation. EXECUTIVE OFFICE OF THE PRESIDENT (1994). Co-ordinating geographic data acquisition and access: the National Spatial Data Infrastructure, Executive Order 12906, Federal Register, 59 17671–17674. GAG (1995). Zwischenbericht zum Projekt Aufbau und Einfürung eines grenzüberschreitenden interregional Telematiknetzes für Geografische Informationssyesteme. Projekt Geographische Informationssysteme, Jahresbericht 1995 (Norden: Gemeinnutzige Ausbildungs—und Entwick-lungsgesellenschaft). GLOVER, J. (1996). OPEN GIS™: removing the barrier for a spatial data infrastrucutre in Proceedings of AGI ‘96, pp. 2.10.1–2.20.5. London: Association for Geographic Information. IRELAND, P. (1995). Data in the fast lane: the race to complete European digital road maps accelerates, GIS Europe, 4(2), 26–7. McKEE, L. (1996). Building the GSDI. Discussion paper prepared for the September 1996 Emerging Global Spatial Data Infrastructure Conference, Bonn, Germany. MORANT, R.C. (1996). Development of a GIS for the Pyrenees work community. Paper presented at the GISDATA specialist meeting Geographic Information: the European Dimension, Buoux, France. MOUNSEY, M. (1991). Multisource, multinational environmental GIS: lessons learnt from CORINE, in MAGUIRE, D., GOODCHILD, M. and RHIND, D. (Eds), Geographical Information Systems: Principles and Practices. London: Longman. RAVI (1995). The National Geographic Information Infrastructure. Amersfoort: RAVI. RIVM (1994). The Preparation of a European Land Use Database, RIVM report no. 712401001. Bilthoven: RIVM. SCHALLER, J., EBERT, T. and VODEMSKY, J. (1995). Three heads are better than one: dissolving national boundaries for effective planning, GIS Europe, 4(8), 28–30. SCHELL, D. (1995). What is the meaning of standards consortia? GIS World, 8(8), 82. ZONNEVELD, W. and D’HONDT, F. (1994). Aménagement du territoire dans la Benelux, Commission spéciale pour l’aménagement du territoire. Brussels: Benelux Economic.

PART ONE

The Creation of Multinational Databases

CHAPTER TWO

From an understanding of European GI economic activity to the reality of a European data set FRANÇOIS SALGÉ

Introduction As the title suggests, this chapter is divided into two parts. The first part considers the nature of the economic sector of geographic information. Many qualitative arguments exist in documents such as GI2000 explaining why investing in the ‘Europeanisation’ of geographic information is so important (CEC, 1996), but very few figures exist showing the size of the sector. Many people think that the GI sector will influence the future of the information technology (IT) field but very few facts (or figures) are available to prove the assertion. The second part of the chapter considers the issue from a rather different perspective. It describes the reality of creating pan-European products. It reports on practical experience in that field and tries to extrapolate the costs of producing a European topographical information template on the basis of this experience. In the final section of this chapter it is argued that such attempts to promote the cooperation of public bodies create dilemmas for those public bodies also involved in offering services. It also voices the user schizophrenia which combines the threat of potential unfair competition from the public bodies with the request for free access to public data (Policy Studies Institute, 1995).

Geographic information economic activity Definition Many people use the word ‘market’ to characterise GI activity. This concept often excludes the activities of governments, either national or local, in creating and using geographic information. For this reason the expression ‘economic activity’ is preferred hereinafter in order to encompass all people, including governments, whose work involves the use of geographic information.

17

18

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 2.1 Estimated size of ‘directly located’ GI economic activity in France (FF million/year)

Geographic information is a buzzword for which no clear definition exists. I will take here the definition currently used within CEN/TC 287: ‘Geographic information is information concerning phenomena directly or indirectly associated with a location relative to the Earth’ (CEN, 1996). This definition is wide enough to encompass any phenomenon, ‘anything that can be perceived as an occurrence or fact by the senses’ (Collins Dictionary of the English Language, 1986), and any way of associating it to a location. Location here is used in a common sense: ‘a site or position, the act or process of locating or the state of being located’ (Collins). The only limitation is the Earth. Basically the direct location is given by coordinates, and the indirect location by a reference to phenomena, themselves linked to a direct location, such as addresses (either physical or postal), NUTS (Eurostat Nomenclature des Unités Territoriales Statistiques) codes or property identifiers.

Estimating the size of EU economic activity for Gl Estimating the size of the geographic information economic activity is impossible with the previous wide definition as there is no economic activity which completely excludes geographic information. Nonetheless, the following figures try to circumvent the issue by considering the geographic information which is only directly associated with a location relative to the Earth. Table 2.1 summarises some of the findings of a study commissioned by the French umbrella organisation for geographic information (CNIG, 1996). From this it can be seen that the directly located GI economic activity in France each year accounts for about FF 12 billion (about ECU 185 million). Nearly half this expenditure is associated with cadastral activities, and the utilities and local government are responsible for another quarter. It has also been estimated that this economic activity provides work for 35,000 people. Given that the total population of the EU is 6.38 times that of France, and making allowance for the fact that GNP per capita in France is rather higher than the EU average, it can be estimated that directly located GI activities may account for at least ECU 10 billion each year and employ at least 200,000 people.

FROM GI ECONOMIC ACTIVITY TO THE REALITY OF A DATA SET

19

Figure 2.1 Geographical location of GIE members.

The reality of a European data set Having provided some considerations on the economic sector of geographic information, the following section discusses the reality of day-to-day work attempting to ‘Européanise’ the national endeavours towards modern digital geographic (topographic) information provision.

MEGRIN and its basic purpose The partners of the GIE MEGRIN (Multipurpose European Ground Related Information Network; see Figure 2.1) are all official mapping agencies of Europe which have decided to invest in the development of the MEGRIN concept. The GIE agreement was signed on 25 September 1995 in Budapest, Hungary. The GIE MEGRIN is governed by the French law on Group de l’Intéret Economique (GIE). It includes contracts with third parties and will initially exist for a period of five years ending in November 2000. The GIE MEGRIN has its registered office in Saint Mandé near Paris (MEGRIN, 1996a). The objectives of GIE MEGRIN are as follows:

¡ to stimulate the market for European GI and facilitate the distribution, in foreign countries, of geographic data produced by the partners;

¡ to facilitate and develop the related activities of its partners; ¡ to improve the image and increase the influence of its partners, both in their own countries and abroad; and

¡ to enable the partners to benefit from work accomplished in common.

20

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

The GIE MEGRIN is committed to working in two main areas:

¡ pragmatic activities aimed at developing better awareness of the availability of European

¡

digital geographic information products and developing merged data sets. The former activity has led to the creation of a system providing information on available data sets (GDDD) which can be accessed through the Internet by any user. The aim is to provide support to that segment of the GI economic sector which seeks to make data about one country available outside that country. The latter activity involves the creation of a seamless data set of the administrative units (SABE) in Europe which can be assigned to any third party who may require a comprehensive knowledge of the territorial organisation of each country; long-term activities, with a view to providing geographical base data (PETIT) in Europe resulting from the merging of the existing databases available from the national mapping agencies. These base data will serve as a topographical reference for geographical information created throughout Europe.

GDDD: a comprehensive metadata service MEGRIN launched its Geographical Data Description Directory (GDDD) project in 1993 to gather information on European digital geographical data. In its present stage the GDDD comprises an implementation of the draft metadata standard on geographic information worked out by CEN/TC 287, the technical committee of the European standardisation body, Comité Européen de Normalisation, in charge of standards in the field of geographic information; (see Larsen, 1996). Metadata are information about data. The term is relatively new and therefore less developed in the context of GI while other branches like libraries have a long tradition of using metadata (book catalogues) to describe their data (books). Metadata on GI is an issue which is regarded with increasing interest from all parties involved in GI systems. The demand for metadata is fuelled by the increasing volume of digital data which are now available from various sources/suppliers for GIS applications. The geographic metadata were acquired from European national mapping agencies who completed a questionnaire designed in accordance with the GDDD metadata model. The questionnaire was sent to 36 national mapping agencies (NMAs) (14 in Germany), who decided which data sets they would present in the GDDD. The response from the national mapping agencies resulted in some 185 data set descriptions being loaded in to the database. High quality data description is essential for a reliable metadata service. Data consistency was checked by analysing the keywords assigned to the data sets. Fortunately, the GDDD contained metadata on several data sets based on the same data model. However, the data sets were described by different people. The analysis revealed inconsistencies introduced by the data providers’ selection of keywords. The disagreement as to the choice of keywords was particularly significant concerning keywords for complex themes such as ‘topography’ and ‘land cover’, whereas the data providers agreed on keywords where interpretation is straightforward such as ‘road network’ and ‘railway network’. This clearly illustrates the importance of having the same understanding of the metadata terminology applied to describe data sets. The keywords were supplied with definitions and the NMAs were asked to check their data set descriptions in order to compensate for the inconsistencies with the use of keywords. This quality check was responsible for a considerable improvement in data description consistency.

FROM GI ECONOMIC ACTIVITY TO THE REALITY OF A DATA SET

21

Figure 2.2 GDDD version 2.0 system implementation.

Today’s technology does not provide database solutions that meet the wide range of the user requirements identified during the pilot project. On one hand, the GDDD should provide a maximum facility in tools for data analysis, and on the other the GDDD should be accessible for untrained users. The system implementation of the GDDD version 2.0 is based on dual technologies: a database implementation and a hypertext implementation. The database solution is used for data analysis and data maintenance. The hypertext solution is distributed on the World Wide Web. Figure 2.2 shows the components of the GDDD version 2.0 system implementation. Geographical metadata are collected from the source, for instance the NMAs, who update national GDDD databases. The national updates are loaded into the GDDD master database where metadata from all data suppliers are stored. Applications are available for production of various outputs for example generation of html files for the Web and for the partial import and export of the master database. The CEN metadata model has been implemented in the central relational database. More than 185 European data sets have been loaded into the GDDD. The html application is available and the GDDD Web pages are automatically generated from the master database. A GDDD service has been established on the Worldwide Web at URL: http:// www.ign.fr/megrin/gddd/gddd.html. A GDDD update tool has been developed. This tool is available to the data providers for updating the national GDDD databases. During the CEN enquiry on the metadata standard, the GDDD will be used as a testbed for the standard. The proposed CEN/TC 287 standards on geographic information are based on conceptual schemata with a high level of abstraction which makes them difficult to implement. The adoption of standards will be encouraged by the availability of implementations. Thus MEGRIN intends to make the GDDD system implementation available on the Internet.

SABE: the product Knowing the current status of available digital topographical national mapping in Europe is just the first step in creating Europe-wide data sets. SABE (Seamless Administrative Boundaries of Europe) is the first of such data sets.

22

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

SABE is a vector dataset of administrative units in Europe (MEGRIN, 1996b). It was created from administrative boundary data provided by NMAs to MEGRIN. The lowlevel administrative units are the units which elect representatives to local councils or, for the EU, NUTS-5 units. As an option, a seamless coastline of Europe database (SCOLE) is available to complement the administrative boundaries which in some countries do not coincide with the actual coastline. SABE contains the following countries: Austria, Belgium, Cyprus, Croatia, Czech Republic, Denmark, Finland, France, Germany, Greece, Great Britain, Hungary, Iceland, Ireland, Italy, Latvia, Liechtenstein, Luxembourg, Netherlands, Northern Ireland, Norway, Portugal, Slovakia, Slovenia, Spain, Sweden, Switzerland. SABE contains the administrative units of Europe at the various administrative levels in each country. The administrative hierarchy of each country is described from the lowlevel administrative unit to the national level. SABE contains for each administrative unit:

¡ ¡ ¡ ¡ ¡

its unique identifier based on the national code; its proper name; its level in the national hierarchy; its level in the NUTS hierarchy, for the EU member states; and for higher-level administrative units, the identification of the lower-level administrative unit.

SABE contains for each administrative boundary its level in the national hierarchy and, for the EU member states, its level in the NUTS hierarchy. SABE is supplied as one layer for each country. International boundaries between two countries are identical in both layers. This means that their geometry (coordinates), topology (nodes and arcs) and semantics (attributes) are the same in both layers. Each low-level administrative unit consists of one or more closed polygons. Each polygon is identified, by an attribute on its centroid, as the main polygon of the administrative unit or an enclave in another unit.

From experience: the tricks of making European data Merging data sets from different national sources involves the resolution of four main types of problem (Illert and Wilski, 1995; Salgé, 1995):

¡ ¡ ¡ ¡

transformation into a unique geodetic referencing system; transformation into a unique geometric and topological frame; transformation into a unique semantic conceptual model; transformation into a unique temporal frame.

CERCO has been involved since the late 1980s in the creation of the European geodetic network (EUREF) based on GPS observation. This has made it possible for the NMAs to provide MEGRIN with the parameters enabling the transformation from the national coordinate system in use (often in a national cartographic projection on a national ellipsoid referring to a national datum) to the geographic coordinate on the GRS 80 ellipsoid referring to the ETRF 89 datum (WGS 84). To some extent this issue is now a known problem at the level of accuracy required for creating European data sets. Creating a unique geometric and topologic frame implies a resolution of the discrepancies which appear at the country borders (see Figure 2.3). Each data set defines the geometry

FROM GI ECONOMIC ACTIVITY TO THE REALITY OF A DATA SET

23

Figure 2.3 Geometric and topological discrepancies between national boundaries.

and topology of international borders. Even if virtually all international borders in Europe are agreed by the parties, discrepancies between combined national data sets exist owing mainly to the accuracy of digitisation. Resolving the geometric discrepancies can be performed by adopting the most accurate geometrical line. Seamless data sets also require full and correct topology. This means that the international borders must contain all the nodes at which the international border is connected to the administrative boundaries of country A and B. Once more the topic is researchable only in the area of making the processes involved more automatic. The semantic issue is more time-consuming. Each country has its own administrative hierarchy: it can be either straightforward, as in Denmark, or very complex, as in Germany (see Table 2.2). Each country provided data sets structured to various levels, from simple drawing files to highly structured databases. The quality of each country’s data set differs: many semantic mistakes were found in the data sets, including administrative units with no geographical extension or geometrical faces linked to many administrative units. The main difficulty is to identify where there is a mistake in the data set and where the reality itself is more complex than the model one has in one’s head: there are pieces of land which belong to several administrative entities even at a national level—for example, Lake Constance is shared between Austria, Germany and Switzerland. The value of a seamless administrative boundaries of Europe data set lies in its ability to be linked to an agreed coding schema such as the NUTS. However, in some countries the perception of the national administrative structure of the National Statistical Institute (NSI) differs from that of the NMA, and the NUTS nomenclature defined by Eurostat also varies over time.

24

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 2.2 Hierarchies of administrative units for selected European countries

FROM GI ECONOMIC ACTIVITY TO THE REALITY OF A DATA SET

25

Figure 2.4 Administrative unit boundaries and water bodies.

Another problem is the question as to whether the area of the administrative unit only includes the land or also includes water bodies (Figure 2.4). This is crucial in coastal areas; for example, do fjords in the Nordic countries belong to the neighbouring administrative units, and are the big inland water bodies (such as Lake Vänern in Sweden or Lake Como in Italy) split between the waterside administrative units? MEGRIN’s solution to this problem is to create a separate layer containing the Seamless Coastline of Europe (SCOLE) which allows for the differentiation of sea and land. The next extension is to incorporate some rough land-cover information giving a first indication of where the water bodies, the built-up areas or the forests are, and it is expected that this will lead to real problems of content definition. These are not specifically European problems but reflect the fact that each country of Europe has selected its own slightly unique approach. The last of the four big issues is the temporal frame. In many respects the administrative unit structure of the European countries is a living body. For some countries, changes occur each year. Others change their system according to electoral cycles. It can also be difficult to define what data should be supplied for a given date. In some countries the situation as of 1/1/yy is available on that day. In other countries it may not become available until April or June or even never. To create a data set such as SABE 91 representing the administrative boundaries as of 1/1/91 which is compatible with the statistical information, it is essential that the data suppliers understand which version of the data they have to deliver. For example, the data required in one country represented the 1984 situation because the 1991 statistics referred to that data set. Still pending is the issue of monitoring the changes through time in order to be able to investigate the evolution of socio-economic patterns of administrative units.

SABE: from statistics to pricing policy Currently SABE is available at two levels of resolution: SABE 30–30 metre resolution, which corresponds to an approximate scale of 1:100,000—and a derived product, SABE 200–200 metre resolution, which corresponds to an approximate scale of 1:1,000,000. Table 2.3 provides some statistics for the currently available SABE 91 products to give an indication of their magnitude.

26

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 2.3 Some key statistics for SABE 91

From Table 2.3 it can be seen that the data set contains details of the boundaries of 105,000 low-level administrative units in 25 European countries. The main difference between the two versions of the database is in terms of the number of points involved. In this respect, S ABE 30 has nearly three times the number of points contained in SABE 200. This is reflected in the overall size of the data set which is 250 Mbytes for SABE 30 as against 165 Mbytes for SABE 200. SABE is now available at an overall cost of less than ECU 10,000 for SABE 200 or ECU 100,000 for SABE 30; annual updates cost 10 per cent of the initial cost. The key principle underlying these prices is the need to ensure that no user can obtain cheaper data for a given country by buying SABE as against the equivalent national data set. Consequently, the overall price of these data sets is governed by the combined price of the equivalent national data sets.

SABE collection of agreements The development of SABE also requires agreements with the data suppliers and with the distributors (Figure 2.5). MEGRIN has not the resources to create and maintain a sales network. This means that SABE products are marketed largely by the sales networks of the NMAs. MEGRIN has an assignment and distribution agreement with all the organisations contributing to SABE which defines, on the one hand, the conditions under which the national data sets are supplied to MEGRIN for the creation of SABE and, on the other hand, the conditions under which the NMAs operate as distributors of SABE. These assignment and distribution agreements are all similar and also include provisions on how MEGRIN can handle direct releases to international organisations such as EUROSTAT. Associated with the assignment and distribution agreements, MEGRIN has developed a contract binding the third party supplied with SABE and the distributing NMA, the latter acting on behalf of MEGRIN. Issues such as ownership, copyright, liabilities and applicable laws are defined within this agreement. In order to increase flexibility in distribution MEGRIN also negotiates with private sector distributors enabling them to add value to SABE and provide new services.

Extrapolation for other topographic themes Creation of EU data sets from scratch Based on information provided by some mapping agencies, it is possible to make an estimate of the costs of creating a Europe-wide 30 metre accuracy thematic data set from

FROM GI ECONOMIC ACTIVITY TO THE REALITY OF A DATA SET

27

Figure 2.5 The SABE collection of agreements.

existing paper maps. Table 2.4 shows that the overall cost of creating national databases for the EU containing communication, river and electricity networks, and land use as well as administrative boundaries is probably of the order of ECU 184 million and that a further ECU 11 million would be required to cover the cost of annual updates. It should be noted that the costs of creating administrative boundaries data sets are low by comparison with those involved in creating data sets for communications networks, river systems and land use.

Creation of EU data sets from existing databases On the basis of the SABE experience it can be estimated that creating an integrated Europewide database might cost another ECU 10 million if data were already available in all countries. However, according to the GDDD, only half the European territory is covered

28

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 2.4 Estimate of the cost of creating 30 metre accuracy national databases for the EU

by digital 30 metre accuracy national databases. Consequently, a further ECU 100 million would be required to create a European topographical information template. From this it can be seen that creating EU data sets from existing databases would cost at least half the price of creating them from scratch.

The dilemma for public bodies One of the major issues concerning the GI activity in Europe is, who is to provide the information on available geographic data in Europe. Metadata services are increasingly available on the Internet or through other types of media, financed mainly by public money. There is also the will of the public sector to promote metadata concepts and standards. It is natural for public bodies, or publicly owned private bodies, to make metadata tools available to the market via the WWW. However, there are private enterprises making their profits from the sale of metadata software who will not be pleased to see such initiatives taken by public bodies. They may even sue the public bodies for unfair competition where the public interest is the main motive of the latter. Pricing data sets assembled from available national data sets (themselves on lease under commercial conditions) illustrates the dilemma facing public bodies trying to build a European information template. The first strategy takes account of the real costs of producing seamless data sets and tries to recover them. This may lead to a pricing policy which proves to be too high for the users. They may even feel that the data producer profits from its monopolistic situation if s/he benefits from any special arrangement with the public sector data owner. An alternative is to set attractive prices to stimulate the market, i.e. set the price at the level which is affordable by the users. Prices might then be divided by 10 or even 100. Then the private sector producer may sue the public body for unfair competition as the latter may not have to recover full costs and may enter the market under relaxed constraints. The difficulty in the GI sector in Europe is to define what is the minimum level of investment by public money which enables the activity to increase and, in the mean time, what are the funding mechanisms which will allow this to happen. Taking the example of the European topographical template referred to above, one could suggest that the Commission should pay the costs of merging the data sets (who will do the work will be determined according to the usual Commission funding rules) and that each country

FROM GI ECONOMIC ACTIVITY TO THE REALITY OF A DATA SET

29

should contribute to the creation of its own national data (according to the subsidiarity principle, it would be up to each country to decide how and who will produce the requested data sets). It may also be suggested that the Commission provides financial assistance to some of the ‘lagging’ countries. However, it should be noted that if there is a requirement that public bodies such as the NMAs are to be given the task of producing the data sets, and if there is also a requirement for the data to be available at low cost, then the data production will be largely subsidised by public money. The main argument for investing in GI is that it provides a better knowledge of our living space in order to make more informed decisions. The economic argument is not thus in terms of a return on investment but rather in increased savings in public expenditure.

NOTE This chapter was written after a presentation given in Buoux (Provence, France) during the GISDATA specialist meeting on the European dimension of geographic information. The ideas expressed in this chapter are the author’s own views and are not those of his organisation nor do they reflect any official statement.

REFERENCES CEC (Commission of the European Communities) (1996). GI2000: Towards a European Policy Framework for Geographic Information. Luxembourg: DGXIII. CEN (1996). Revised text of prEN 12009, Geographic Information—Reference model, CEN/TC 287 N461. Brussels: Comité Européen de Normalisation. CNIG (1996). Internal documentation of the pilot committee of the Coopers & Lybrand study. Paris: Conseil National de l’Information Géographique. ILLERT, A. and WILSKI, I. (1995). Integration and harmonisation of contributions to a European dataset. Presentation to 17th International Cartographic Association Conference, Barcelona, Spain. LARSEN, P.L. (1996). Geographical Data Description Directory: an implementation of the CEN/ TC 287 metadata standard on geographic information, GIS Europe, 5(7), 20–2. MEGREM (1996a). http://www.ign.fr/megrin/megrin.html. MEGRIN (1996b). SABE 91 Infopack. Saint Mandé: Multipurpose European Ground Related Information Network. POLICY STUDIES INSTITUTE (1995). Publaw III, final report prepared by PSI (London) and GRID (Namur). London: Policy Studies Institute. SALGÉ, F. (1995). Intégration de bases de données géographiques hétérogènes: le cas desdonnées topographiques, Journées d’étude bases de données et systèmes d’information pour l’environnement. Versailles: Institut National de la Recherche en Informatique et Automatique.

CHAPTER THREE

Transboundary European GIS databases: Review of the Baltic region experiences SINDRE LANGAAS

Introduction The number of seamless GIS databases for European transborder regions is growing at a rapid pace (CEC, 1996; Masser and Salgé, 1996). These developments have, until recently, taken place despite a generally unfavourable European ‘climate’ with respect to the institutional, legal, commercial and technical problems in assembling GIS data from several countries. The new trends are driven by two factors, as pointed out in the recent European geographic information infrastructure debate (CEC, 1996). The first is the escalating internationalisation in most sectors of society, especially in the business, governmental and environmental sectors. The expansion of the ‘geographical arenas’ in which processes and activities take place requires transborder geographic information to perform rational analyses and assessments. The second factor is the general growth in the information technology (IT) business. This expansion, which to a large extent has been focused upon PC users, has led to increased opportunities for supply of cheap, powerful information and communications technology. For example, the Windows 95 version of Microsoft Excel now has in-built a subset of the desktop Maplnfo GIS software. Within this European context this chapter describes past and current attempts to build seamless multilayer GIS databases of the Baltic Sea region. This region has a long tradition of environmental cooperation, quite notably also during the Cold War (Westing, 1989). More recently, following the disintegration of the Soviet Union with the associated regaining of independence for the three Baltic States and the unification of Germany, cooperation has deepened and expanded into other areas of the public and commercial sectors. A major change has been in the openness by which information can be ex-changed across national boundaries. As a consequence of these profound political changes, a number of Baltic GIS data initiatives appeared in the early 1990s, aiming towards seamless GIS data sets covering the Baltic region. This chapter begins by describing and defining the Baltic region and the needs for seamless Baltic region GIS data sets. Particular emphasis is given to water quality definition in the region as this issue is a key argument for establishing seamless databases. The diverse geopolitical context is also briefly touched upon as it is atypical for the usually 31

32

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 3.1 Status of the main regional GIS database initiatives for the Baltic region.

EU-centric Geographic Information Infrastructures debate. The technological differences and differences in geographic information and data policies among the countries are then taken up. The main part of the chapter describes and examines the development of a number of regional database initiatives shown in Figure 3.1. Special attention will be given to the GIS database originating from the Baltic Drainage Basin Project.

The Baltic region: need for seamless GIS data and related issues The need for Baltic region GIS databases can best be appreciated by considering various ‘actor’ groups in the Baltic region and their requirements for spatial information to support their activities. Until quite recently, the main actor group was the environment, hydrometeorology and marine science group. Lately, other groups have become more active, groups that begin to or most probably soon will request Baltic region spatial data sets. These are actors in the spatial (physical) planning, the political and the business communities. The Baltic Sea States Summit on the island of Gotland in May 1996 strongly indicated the upcoming needs.

Environmental, hydrometeorological and marine science needs The Helsinki Convention of 1974, founded to protect the marine environment of the Baltic Sea, was the first international agreement to cover all sources of marine pollution,

TRANSBOUNDARY EUROPEAN GIS DATABASES

33

both from land and from ships as well as airborne. To accomplish its aim, the Convention called for action to curb various sources of pollution. In this context the Helsinki Commission (HELCOM) Baltic monitoring programme, which supports environmental assessment activities, has increasingly started to ask for and use GIS data in its work. The Baltic Sea Joint Comprehensive Environmental Action Programme (JCP) was developed and approved in 1992 to ‘assure the ecological restoration of the Baltic Sea’. The JCP is coordinated by the Helsinki Commission and is being implemented over twenty years (1993–2012) at an estimated total cost of ECU 18 billion. Its major focus is to reduce point and non-point sources of pollution, particularly at the 132 (now 124) ‘hot spots’ identified by the JCP. The hot spots are mostly municipal waste and industrial sites. Despite an obvious need (at least to GIS persons) for a spatially driven information system to measure progress and achievements, no strong demand has yet come from the coordinators. Scientific Baltic Sea regional enterprises, with thematically similar but in many cases more analytical approaches, have increasingly taken GIS into use. The Baltic Drainage Basin Project (BDBP) and the Baltic Sea Experiment (BALTEX) projects are prime examples of this, both demanding and using spatial data sets. The overall objective of the BDBP was to find sustainable development patterns where the significance of the Baltic region’s natural resource base is taken into account and efficiently managed in both an economic and ecological sense. One of the project’s main foci was the eutrophication of the Baltic Sea and its causes and effects (Gren et al., 1996; Sweitzer et al., 1996). BALTEX seeks to achieve the following:

¡ to explore and model the various mechanisms determining the space and time variability ¡ ¡

of energy and water budgets of the BALTEX area and their interactions with surrounding regions; to relate these mechanisms to the large-scale circulation systems in the atmosphere and oceans over the globe; and to develop transportable methodologies in order to contribute to the basic needs of climate-, climate impact- and environmental research in other regions of the world (Raschke, 1994).

It should be noted from these two examples that the delineation of the ‘functional’ Baltic region varies. Owing to the meteorological modelling component, data from a large buffer zone outside the drainage basin is needed for BALTEX purposes (see Plate 1).

Spatial planning and political needs The environmental, hydrometeorological and marine science needs for GIS data have increased substantially over the past ten years. On the other hand, the physical planning community was quiet until 1992. Then, the Vision and Strategies for the Baltic Region (VASAB) 2010 project was initiated by the Baltic region ministers of physical planning. During the development of the first VASAB reports, the lack of spatial data sets showing inter alia communications infrastructure was identified. Plate 2 shows a map from the VASAB 2010 report and the area defined as the Baltic region according to this perspective. This definition is political and not restricted to the drainage area. The MapBSR database initiative described later is a direct response to the scarcity of physical-planning-relevant GIS data sets.

34

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

The reinforced high-level focus upon the Baltic region, demonstrated by the recent Baltic Sea Summit on Gotland, Sweden, in May 1996, points clearly to the growing need for comparable socio-economic data. These data sets are needed to support political agenda setting and policy developments on various political/administrative levels. Whether these needs will be met by the MapBSR project, or whether the GISCO project of EUROSTAT will be expanded to cover also the European continent remains to be seen.

Business needs Although it has not yet been explicitly expressed, there is certainly going to be an increasing need for GIS data in support of Baltic regional business activities. In particular, multinational companies looking upon the Baltic region as a separate sales area will demand marketrelevant and comparable spatial information for the region. Equally, databases for in-car navigation will certainly be demanded for the entire Baltic region.

Delineation of the Baltic Sea region In the case of GIS database building for the Baltic Sea region, most of the interest so far has centred around water quality pollution in the Baltic Sea itself. This interest directly tends to limit the area to the Baltic Sea drainage basin. However, as argued in Gren et al. (1996) and mentioned earlier in the context of the BALTEX project, when considering, for example, the sources of atmospheric nitrogen pollution, the functional area of interest has also to include the areas in which the air pollution originates. As a result, the political definition of the Baltic region varies according to the regional cooperation bodies and initiatives. In some definitions, based upon cooperation between nation states, the Baltic region extends from Greenland in the West to the eastern part of Russia. This fuzziness in the delineation of the Baltic Sea region is a key point in a recent review of Baltic regional cooperation (Stålvant, 1996). The main section of this chapter deals with Baltic Sea region defined in terms of the drainage basin. In this case, the region includes all or part of fourteen countries: Finland, Russia, Belarus, Estonia, Latvia, Lithuania, Ukraine, Slovakia, Czech Republic, Poland, Germany, Denmark, Norway and Sweden. Geopolitically, this is a diverse mix of European countries with, for example, four EU and three GIS countries. The technological and economic gap between these countries is huge. This is also a factor of importance in the development of GIS databases. The total area of the Baltic Sea region, using the drainage basin definition, is 1,745,000 km2. About 85,000,000 people live within the area.

The first generation complete GIS database of the Baltic region The first easily accessible multi-thematic GIS database with complete coverage of the entire Baltic region was created in a very simple manner (Langaas, 1992a,b). In 1992, GRID-Arendal, a centre in the United Nations Environment Programme’s (UNEP) network of collaborating Global Resource Information Database (GRID) environmental data and information centres, created a small-scale or coarse-resolution database of the Baltic Sea region using a series of global GRID data sets. At the time of this initiative, no easily accessible seamless GIS data sets of the entire Baltic drainage basin existed. The only data sets covering the entire region available to the

TRANSBOUNDARY EUROPEAN GIS DATABASES

35

GIS community at that time were the global GRID data sets available from GRIDGeneva which had been created at research institutions or other UN bodies. GRID provided the GIS data sets for further dissemination to other researchers, decision-makers and educational institutions but imposed restrictions on commercial users (see Chapter 8). The data medium used by GRID-Geneva at that time to disseminate GIS data sets was computer compatible tapes (CCT). This medium was not very widely used in the Baltic Sea region, and hence very few institutions had access to these data. To improve the accessibility to these data sets, GRID-Arendal decided to extract the Nordic/ Baltic region from about twenty of the global GRID data sets and distribute them as a small package of four floppy disks with an associated data document. The intended application areas were primarily GIS education, research and planning. This initiative was well received and fulfilled a regional need (Langaas, 1994). Because most data sets were initially global in scope, the applications of the database were found mainly in the GIS educational sector. Still, some applications of the vector part of the database (i.e. subsets of CIA World Databank II) involved cartographic visualisation and more advanced GIS techniques (see, for example, Lucas and Bivand, 1994). This GIS database and the documentation was also made available in 1995 through the Internet at http://www.grida.no/nor-balt/index.htm. The inherent characteristics of this database—its small scale, low resolution, inconsistency, lack of (environmentally and otherwise) relevant topics—made subsequent database developments imperative.

The second generation complete GIS database of the Baltic region A consortium of several research groups in Sweden, Norway, Poland, Germany, and the UK carried out a joint research project, the Baltic Drainage Basin Project (BDBP), in 1993/94 under the EU Environmental Research Programme of 1991–94. The overall aims of the project were:

¡ to assess the linkages between economy and ecology related to the nutrient pollution to the Baltic Sea;

¡ to evaluate various ecological, technological, economical and institutional instruments to reduce the pollution; and

¡ to assess the ecological effects of reduced nutrient pollution. In subproject 1 (land use and ecological carrying capacity of the drainage basin), a research group consisting of the Beijer Institute and the Department of Systems Ecology, Stockholm University, and GRID-Arendal, Norway, developed a seamless, multi-thematic and coherent GIS database for the drainage basin. Emphasis was given to land cover and population. The main applications of this GIS database in the BDBP were:

¡ to obtain reliable statistics on land cover, population and country distribution within the overall drainage basin as well as for sub-basins (Sweitzer et al., 1996); and

¡ to apply it to more focused research (see, for example, Gren et al., 1996; Jansson et al., 1996). For a comprehensive article on the technical and methodological aspects related to the development of the database, and some applications of it, the reader is referred to Sweitzer et al. (1996). The project group decided to make a slightly modified version of the GIS database available to the public on the Internet. The development of this database was carried out

36

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

after the BDBP formally ended with complementary funds from the Nordic Council of Ministers and GRID-Arendal. This GIS database, together with ready-made cartographic products in standard graphical formats, was released on the Internet in August 1995. This GIS database is one of the rare examples of a publicly available online international GIS database in Europe. The main purpose of this section is to highlight some experiences in the compilation of the database. Most of the primary data sources used for the development were taken from the public domain, such as the Digital Chart of the World and the European Space Agency’s Remote Sensing Forest Map of Europe. Therefore, the focus will be put on the administrative boundaries data set, which from the legal and institutional perspectives are the most interesting aspects to discuss.

Data compilation and editing for research Based upon extensive knowledge of existing sources for administrative boundaries, a pragmatic approach was taken in the assembling of data sets to be used in the research. The adoption of this approach was dictated by a limited budget, which in some cases prevented the best possible and more costly sources being used. The intended main purpose of the administrative units data set was to model population density by linking it with population statistics and land cover data. Therefore, the administrative unit level sought for by the application was the highest (most detailed) for which population statistics could readily be obtained. For Denmark, Finland, Germany, Norway and Sweden this meant second-order units (NUTS level III). First-order subnational units were found appropriate for the remaining countries. The nominal scale for the final multi-thematic database was 1:1,000,000, and data sources around this scale were sought. The idea was to release the database in the public domain. Therefore, the legal issues surrounding redistribution were also considered. We were not too concerned with aspects like positional accuracy: 1990 was used as a nominal base year; in practice, however, the actual years represent a range from 1985 to 1993. An overview of the data sources or the providers of the administrative unit GIS data is given in Table 3.1. All these data sets, except the Lithuanian, existed already in digital format. The data sources most easy accessible were the ones existing on CD-ROMs (ArcWorld), and available on Internet (EpiMap boundary files). UNEP-collaborating partners in Estonia (Estonian Environment Information Centre), Latvia (Latvian Environment Data Centre), Poland (GRID-Warsaw) and Switzerland (GRID-Geneva) kindly provided the remaining data sets. The data set covering the Nordic countries, the Nordic Cartographic Database (NCD), was provided by the Nordic Mapping Agencies as a beta version before they had finalised the NCD and determined its price and release policy (see Chapter 10). The various GIS data sets were concatenated using the Political and Oceans Layer (PONET) of the Digital Chart the World (DCW) as a common ‘glue’ data set for all international boundaries and coastlines. The concatenation required that all the various data sets be converted into pcArc/Info vector format and georectified to a common Lambert Azimuthal Equal Area projection. The use of the PONET land/sea boundary in all other layers in the BDBP database ensured consistency in land/sea demarcation. Population statistics were compiled in parallel with the assembling of boundary GIS data files. The year of the population statistics was different from the year of the boundary files for many countries, so adjustment of the boundary files was sometimes necessary.

TRANSBOUNDARY EUROPEAN GIS DATABASES

37

Table 3.1 Data sources and approximate nominal scales of the BDBP administrative units data set

Finally, the vector database was rasterised into 1 km2 cell size and converted into Idrisi raster format for analysis. Population statistics for the various administrative units were obtained from several sources, mostly statistical yearbooks. The four Nordic countries were represented by their second-order administrative units. Statistics for these units were bought from the national statistical agencies (NSAs). These payments were related to handling costs rather than data costs. Generally speaking, the identification of existing data sources proved not to be too difficult. This, was, however, more related to some years of international GIS experience and cooperation in the Baltic region than to the availability of an easily accessible data directory (or metadata) system. The actual acquisition of data sets worked well in most cases. No data set (except the statistics for the Nordic countries) had to be bought. The data sets were either made available free of charge, were already in our possession, or were downloaded from the Internet. In some cases the issue of copyright and right to redistribution was raised as a concern by institutions providing data since it was explicitly mentioned that the intention was to publish the data in the public domain. Detailed metadata provided by the data providers were generally limited or missing. For example, data set lineage descriptions were in almost all cases missing.

Distribution of the BDBP database on the Internet The distribution of the seamless BDBP GIS database based upon a wide variety of primary data sources needed thorough consideration. Some data sources were not controversial to redistribute; other data sets were more questionable. Examples of the former were the boundary files based upon the ArcWorld, EpiMap and the data set provided by the Estonian Environment Information Centre. These were presumed to be or were explicitly placed in the public domain. The Nordic Cartographic Database (NCD) and the EUROSTAT/ GISCO and GRID-Geneva data set belonged to the questionable category. The NCD was developed jointly by the national mapping agencies (NMAs) in the five Nordic countries

38

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

(Persson, 1995). Given the fact that GIS data sets developed by the Nordic NMAs are usually embedded in copyright restrictions, so too, presumably, is this data set. The data set given as EUROSTAT and GRID-Geneva actually originates from EUROSTAT but has been further improved and corrected by GRID-Geneva. GRID-Geneva had requested EUROSTAT to redistribute this data set in an edited and corrected form. This request was not granted owing to the uncertainties related to the future distribution and pricing policy of EUROSTAT. The statistical population data associated with the administrative units were intended to be distributed with the administrative units boundary files as attribute data. These associated attribute data further complicated matters. Practical rules of conduct and legal practices are rarely provided for how the redistribution of assembled international GIS data sets derived from small-scale and coarse-resolution GIS data sets originating from various sources can and should be performed. Therefore, a sensible solution was designed which we considered would be acceptable to the original data providers. We decided to distribute the database in both vector and raster format—in pcArc/Info and Idrisi formats, respectively. Generally, we anticipated that the rasterised version of the administrative units boundary file would be less difficult to redistribute from a copyright perspective since the rasterisation was a data degradation that would mean irreversibility relative to the original data. Therefore, for the vector data we used the same logic. We first rasterised the vector data into a 500 metre resolution, and subsequently vectorised it back again. The population attribute data were also degraded. Instead of a precision of 1, we rounded the data to a precision of 100. In addition, metadata information was created for each data set made using a form based on a mix of UNEP/GRID’s internal metadata tool and DIF. The BDBP database was published on the Internet in August 1995 at http:// www.grida.no/baltic/. The database home page received around 13,000 visits, and more than 2,500 GIS data files were downloaded in the nine-month period from August 1995 to April 1996. In addition, a number of ready-made cartographic products in standard graphical file formats (GIF, TIF and EPS) were created. Later, statistical files were published, developed directly from the GIS database.

Current initiatives Basic Geographic Information of the Baltic Drainage Basin (BGIS) initiative The first initiative to compile a multi-thematic Baltic region seamless database following the breakdown of the former Soviet Union and the German unification, was the Basic Geographic Information of the Baltic Drainage Basin (BGIS) initiative. The Statistical Division of the UN Economic Commission of Europe (ECE) was behind this initiative. It had been given the task by the pan-European Conference of European Statisticians to work towards a drainage-area-based statistical database for the Baltic Sea region. Since it lacked GIS competence, UN ECE hosted a meeting in 1991 in Geneva with some international and national organisations to discuss the establishment of a Baltic GIS. Several talks followed, the initiative matured and in 1994 a feasibility study for the establishment of Basic Geographic Information of the Baltic Drainage Basin (BGIS) database was carried out under the auspices of the Helsinki Commission (RELCOM). This was funded by the Nordic Council of Ministers and carried out by the Finnish

TRANSBOUNDARY EUROPEAN GIS DATABASES

39

Environment Data Centre (now the Finnish Environment Institute), the Swedish Meteorological and Hydrological Institute, the Swedish Space Corporation, GRID-Warsaw and GRID-Arendal; and the BGIS Project Group. The feasibility study recommended a multipurpose (although with an environmental focus) and multi-thematic database based primarily on existing data sources. Thus, the proposal was in many ways similar to the BDBP. However, the ambitions were higher both in terms of scale and resolution and more notably the choice of themes. Owing to difficulties in fundraising for this initiative, the BGIS initiative has been put temporarily on hold. Some of the recommended themes in the BGIS feasibility study have been included in a follow-up proposal for a RELCOM GIS. The HELCOM GIS proposal directly addresses the needs of the Helsinki Commission for a spatial decision support system for HELCOM assessment activities, and more importantly to monitor and evaluate the progress made in the ECU 18 billion Joint Comprehensive Environmental Action Programme. Funding is currently being sought for this proposal. Even though many of the recommended themes have had a cartographical nature, the BGIS and HELCOM GIS initiatives had a clear environmental application focus. The leadership was taken by prominent environmental data and information institutions with considerable GIS experience. Notably, the NMAs, which one would expect to be the first to develop seamless GIS databases of this kind, were more or less absent during the discussions and the open meeting leading to the BGIS feasibility study. This changed with the advent of the MapBSR project.

Map Data Sets in the Baltic Sea Region (MapBSR) The MapBSR project can be looked upon as the response of the NMAs in the international Baltic GIS arena to the other initiatives. The aim of the MapBSR project is to provide uniform, reliable map data sets for the Baltic Sea catchment area and the countries within its sphere of influence. The cartographic database is intended to promote various forms of international cooperation, for instance in the environmental protection of the Baltic Sea. The uses for the database will include natural resource management, environmental impact assessment, administration, transboundary planning, research and map production. The political context of the MapBSR project is the Vision and Strategies around the Baltic Sea (VASAB) 2010 project. Initiated by the ministers of physical planning in 1992 the VASAB project aims at a shared planning vision and strategy for the Baltic region. In 1994 and 1995 two expert seminars on the creation of spatial data sets in the Baltic Sea region were organised by the National Land Surveys of Sweden and Finland, respectively, under the VASAB framework. These provided the background for the initiation of the MapBSR project. The MapBSR project is led by the National Land Survey of Finland and comprises only NMAs and national surveying institutions. The project is still in its initial phase. A feasibility study of the project will be presented at the meeting of the heads of NMAs of the countries around the Baltic Sea in October 1996. The implementation of the project will be decided at the same meeting and close cooperation will be sought with other GIS data organisations. The cartographic database is intended to include cartographic information at a scale of 1:1,000,000 (i.e. coastline, administrative boundaries, rivers and lakes, roads, settlements, altitudes, etc.). The NMAs of each of the participating countries will produce the agreed map elements for the areas of their respective countries. These will then be combined into one cartographic database. Common standards and classifications will also be created, and other pan-

40

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

European projects will be taken into account. In many ways, the MapBSR project can be looked upon as a regional Multipurpose European Ground Related Information Network (see previous chapter).

Conclusion The review of various initiatives clearly shows that there is a great interest in the development of seamless multi-thematic GIS databases for the Baltic Sea region. Many institutions have been and are involved in efforts aiming at improved databases. So far, only two multi-thematic seamless Baltic Sea region databases have been developed. The efforts leading to the creation of these databases can be described by a number of characteristics:

¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡

an environmental application focus small project groups limited budgets pragmatic scientific and technical approach (‘best available information’ philosophy) a raster data focus a limited number of themes small scale and coarse resolution public domain data (input and output)

Other efforts, the BGIS and the MapBSR project, are in many ways more ambitious. They can be characterised by:

¡ ¡ ¡ ¡ ¡ ¡ ¡ ¡

a multi-purpose application focus large consortia involved undefined, but large budgets standardised but still somewhat pragmatic approach a vector data focus a large number of themes public domain data (output only) associated training activities

From these characteristics it is possible to identify some issues that may be critical in development of seamless GIS databases of the Baltic region and possibly elsewhere in Europe.

Database applications The Baltic region experiences indicate that focused efforts driven by one or a few applications are more likely to succeed than multipurpose efforts. This cannot be interpreted very strictly since there are few case examples, but at least the clearly focused BDBP project showed that it is possible to come up with a small and quite coherent multi-thematic database of acceptable quality for regional analysis within a limited time period and with a limited budget.

Database developer consortia It appears that the institutional solution and the size of the group or consortium developing the database has some relevance for the rate of success and, more importantly, for the

TRANSBOUNDARY EUROPEAN GIS DATABASES

41

speed of development. This issue is also connected to the degree of application focus when developing the database. The larger the consortium and the more scattered the intended database applications are, the more difficult it appears to be to obtain tangible results. This is particularly the case for a region like the Baltic, where considerable differences exist among the countries in terms of technological and economic capacity. For consortia that are made up of national representatives it is important that the shared regional database vision outweighs the national (institutional) interests of the consortia members. As pointed out in the GI2000 document (CEC, 1996), it is also an advantage if an international GIS body with a true mandate to work on the international arena in question can be involved.

Data sources, copyright and pricing The choice of quality and detail of primary data sources when compiling a transboundary GIS database for the Baltic region has a notable impact upon the possible obstacles in the compilation. The four main initiatives described have had an increasingly ambitious approach where the primary data sources are concerned. The first generation Baltic GIS database was based upon global, and obviously coarse, GRID data sets that were found entirely in the public domain. The BDBP database was more ambitious and used a mix of global (Digital Chart of the World, Arc World), European (European Space Agency’s Remote Sensing Forest Map of Europe, Nordic Cartographic Database) and national data sources. For the redistribution of seamless data sets created, considerable modifications were made to the seamless data sets considered critical from the copyright point of view. It is recognised that the modifications done may not be sufficient from the point of view of the Berne Convention, but they proved sufficient to satisfy the data providers. The BGIS plans were even more ambitious in terms of quality and detail. The BGIS consortium recognised the problems with data copyright and pricing among the NMAs, and therefore recommended that primary data sources should, to the largest extent possible, be found among public domain data sets (BGIS, 1994). Such problems may be overcome if the database development is for one particular application within one institution. However, when the purpose is to redistribute the data free of charge, the likelihood that the primary data institutions will be more negative increases. The MapBSR project is clearly the most ambitious in this respect. It states explicitly that the idea is to rely entirely upon national data sets. Thereby, this project in many ways resembles a regional Multipurpose European Ground Related Information Network (MEGRIN). MEGRIN makes geographic information from the NMAs of Europe available to users of pan-European geographical data and has been set up with the NMAs as the owner. The MEGRIN experience with the development of the Seamless Administrative Boundaries of Europe (SABE) database (see previous chapter) and national studies (Statskontoret, 1992), clearly forewarn that, under the current European copyright and pricing regime, the MapBSR consortium will have great difficulties in agreeing upon the distribution policies and mechanism. The ambition of the MapBSR project that: Official map and geographic information [created by the MapBSR project] shall be made freely available at the regional level, and not be used for commercial purposes may be difficult to enforce, especially if the intention is that its content will be continuously updated.

42

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

In general, we find the Baltic cases of transboundary seamless GIS database development for public release are part of the ‘cocktail’ of legal and institutional issues related to the creation and distribution of multinational European GIS databases, as nicely phrased by David Rhind (1995). It is quite clear that flexible and inexpensive mechanisms to resolve such issues need to be invented before enterprises of this kind will become legally and institutionally painless. The ambitions of GI2000 (CEC, 1996) in this direction are crucial and clearly need to be supported by all involved European GIS parties in order to become reality.

Financial resources A most critical success factor, however, for the enterprises described for the Baltic region has been the financial dimension. The financial aspect are related to all the issues discussed earlier. The development of transboundary GIS databases will depend on existing data sources. Therefore, one may anticipate that this is inexpensive since no primary data collection is necessary. However, the experiences from the Baltic region, and elsewhere in Europe, have shown that the costs involved are considerable. In most cases these costs are related to the working costs incurred in solving the many issues involved in the establishment of European transboundary GIS databases. These are related to the time needed for the development of harmonised data models and standards. Nationally developed data sets use various national standards and data models. They require a common set of data models and standards to be used for a common region. Common data models and standards take time to develop and require resources: and the larger the consortia involved, the higher the costs. However, global and European standardisation efforts under the auspices of the International Organisation for Standardisation (ISO) and the European Committee of Standardisation (CEN) may reduce such work in the future. The actual compilation work at the front of the GIS also generally requires more resources than one would expect. The reasons for this are several: different national projection systems are used, there are many GIS data formats, there is a lack of detailed metadata associated with the individual data sets, and frequently data quality is suboptimal. The ‘transactional’ costs involved in resolving legal and institutional issues related to transboundary GIS databases in Europe are probably a prime reason why the only panEuropean GIS database success stories cited in the GI2000 document are the commercial road network databases. They are obviously of high commercial value and can thus make up for such transactional costs. For other, less wealthy GIS application areas, such as international environmental management, high transactional costs can be critical. In conclusion, the experiences from the Baltic region are by no means unique in a European context. Actually, the Baltic region is, relatively, in a quite good position compared with many other European transboundary regions mentioned in the GI2000 document. This is owing mainly to the fact that the region, here defined as the watershed region, is a transboundary hydro-ecological region sharing an environmental problem: the state of the environment of the Baltic Sea. This has led to a number of initiatives from within the environmental GIS community. The tangible results so far have been two multi-thematic small-scale GIS databases. The Baltic Drainage Basin Project GIS database is possibly the most comprehensive publicly available transboundary European GIS database on the Internet. Two other initiatives aiming at seamless databases of higher quality, with more themes and at a larger scale (the BGIS and MapBSR projects) have gone on for a number of years. The untangling of the various issues identified earlier will determine whether or when the aspirations of these two initiatives will materialise.

TRANSBOUNDARY EUROPEAN GIS DATABASES

43

REFERENCES BGIS (1994). The Basic Geographic Information of the Baltic Drainage Basin. Helsinki: National Board of Waters and the Environment. CEC (Commission of the European Communities) (1996). GI2000: Towards a European Policy Framework for Geographic Information. Luxembourg: DGXIII. GREN, I-M., SÖDERQVIST, T., WULFF, F., LANGAAS, S., SANDSTRÖM, M. and FOLKE, C. (1996). Reduced nutrient loads to the Baltic Sea: Ecological consequences, costs and benefits, Beijer Discussion Paper Series No. 83, Stockholm: Beijer International Institute of Ecological Economics. JANSSON, A., FOLKE, C. and LANGAAS, S. (1996). Quantifying the nitrogen retention capacity of natural wetlands in the large-scale drainage basin of the Baltic Sea, (Submitted to International Journal of Landscape Ecology). LANGAAS, S. (Ed.) (1992a). Regional environmental GIS data sets of the Nordic countries and the Baltic Sea drainage basin, GRID-Arendal Report Series No. 1. LANGAAS, S. (1992b). Environmental GIS data for education, research and planning covering the Nordic countries and the Baltic Sea region, Kart og Plan, 52(4), 287–9. LANGAAS, S. (1994). Global GIS data made regional, Proceedings GIS Baltic Sea States ‘93, Tallinn, Estonia, 29 November-1 December 1993, pp. 169–76. LUCAS, A. and BIVAND, R. (1994). Inference, data interpretation and GIS: Comparative studies with a GRID data set, Discussion Paper No. 200, Dept of Geography, University of Bergen. MASSER, I. and SALGÉ, F. (1996). The European Geographic Information Infrastructure debate, in Craglia, M. and Couclelis, H. (Eds), Geographic Information Research: Bridging the Atlantic, pp. 28– 36. London: Taylor and Francis. PERSSON, I. (1995). Examples and experiences from some international joint projects. Kartbladet, 4, 40–3 (in Swedish). RASCHKE, E. (Ed.) (1994). Scientific Plan for the Baltic Sea Experiment (BALTEX), 2nd edition. Geestchaht: GKSS Research Center. RHIND, D. (1995). Spatial databases and information policy: A British perspective, Proceedings of the Conference on Law and Information Policy for Spatial Databases, October 28–29, 1994, Tempe, AZ, NCGIA, University of Maine, Orono, pp. 82–92. STATSKONTORET (1992). Financing of Geographical Data, Report 1992:34. Stockholm: Statskontoret (in Swedish). STÅLVANT, C.E. (1996). Actors around the Baltic Sea. An Inventory of Infrastructures: Initiatives, Agreements and Actors. Prepared for the Baltic Sea States Summit 1996. Stockholm: Ministry of Foreign Affairs. SWEITZER, J., LANGAAS, S. and FOLKE, C. (1996). Land cover and population density in the Baltic Sea drainage basin: A GIS database, Ambio, 25(3), 191–8. WESTING, A. (Ed.). 1989. Comprehensive Security for the Baltic: An Environmental Approach. London: Sage.

PART TWO

Multinational Databases for Specific Applications

CHAPTER FOUR

Development of a GIS for hydrological modelling of the River Rhine JAAP KWADIJK AND ERIK SPROKKEREEF

Introduction In the past decade, interest in computer models that can be used for quantitative assessment of the impact of environmental change on runoff, water availability and quality has grown rapidly. These models are expected to produce reliable results for large regions that range from countries to continents. Modelling of hydrological and environmental processes on such scales requires spatially distributed data of differing nature. Data requirements and model structure strongly depend on the purpose of the model and on the manner in which the developers want to describe the relevant processes. From the policy-maker’s point of view, such models should use and produce variables that are related to socio-economic development. If these models are used to assess the impact of environmental change, such as climatic change or land use change, the natural scientist that develops the model prefers to describe the natural processes on a physical basis. The reason is that when limiting natural conditions, such as precipitation or temperature, change, the reliability of results produced by empirical models will be reduced. Apart from our limited knowledge of many relevant processes, it is for practical reasons not possible to get all the data necessary to develop a model that will satisfy all interested parties. Therefore, such models will typically be based on limited data resources, and will use simplified descriptions of complex natural, economic and social processes. These considerations formed the basis of a study of the impact of climatic change on the discharge of the River Rhine (Kwadijk, 1993). For this purpose a GIS-based water balance model (RHINEFLOW) was developed. The model will be upgraded in the near future because the water management authorities are planning to use it as a tool to assess the impact of climatic change on the discharge of the Rhine and the resulting effects on different river functions such as navigation, drinking water availability and safety. This chapter gives a picture of the methods that were used to collect the data for the development of the first model version. It also describes the strategy for data collection that is being followed to improve the model in the near future. The discussion focuses on the data required for the development of the hydrological model. It does not aim to produce a blueprint of data collection strategies that can be used for research that extends across 47

48

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 4.1 Flow diagram of the RHINEFLOW 1.0 model.

national borders. What it shows, however, is that the needs for data may very well change even during one project and that, as a result, data collection strategies also change. This implies that a very flexible structure for the design of a European geographical database should be chosen.

Development and data collection for RHINEFLOW version 1.0 The RHINEFLOW 1.0 model was developed within the framework of a project started in 1988 at the University of Utrecht, that aimed at assessing the impact of climatic change on the River Rhine. An inventory of available data and models showed that, at that time, no suitable model existed that could be used as a tool for this purpose. Therefore we decided to build such a model. A water balance model of the River Rhine considers the Rhine catchment as a closed basin with one outlet, the River Rhine at the German-Dutch border. Within this basin the model calculates water supply, storage and losses along the path from precipitation to river discharge. The supply of water comes from precipitation that may fall in the form of rain or snow on the soil. Water is stored in the soil from where it may evaporate, which is considered a water loss, it may run off directly to the river or it may percolate to deeper layers. From these deeper layers water flows slowly to the river. Figure 4.1 shows the w ater flow between different components. The amount of losses, storage and flows depends on geographical and climatological characteristics. Precipitation will fall as snow when the air temperature is below 0º Celsius; more water evaporates from forests than from bare ground, and soil types that consist of clay can store more water than sandy soils.

A GIS FOR HYDROLOGICAL MODELLING OF THE RHINE

49

For these reasons a model that simulates runoff and water availability of the River Rhine necessarily employs variables such as land use, soil type, elevation, precipitation and temperature. As these variables vary strongly within the basin, their spatial distribution should be taken into account. The Rhine basin is probably one of the best monitored drainage basins in the world and many databases containing various environmental, meteorological, hydrological and geographical information exist. However, to locate all these databases and determine which data they contain is not an easy task. Moreover, a great number of organisations administer these data for their own (non-hydrological) purposes. Classifications used will serve these purposes, meaning that these data cannot be used directly in hydrological models. Therefore, it was not clear initially what kind of data and what spatial and temporal resolution would become available within the time span of the project (four years). As we had limited financial resources we were certainly not able to buy all the data. Crucial for the model development in this stage was a publication of the International Commission for the Hydrology of the Rhine basin (CHR/KHR). The members of this commission are representatives of the national institutes of inland water management of the riparian states of the River Rhine. The task of the CHR is to initiate and to support research relevant to the hydrology of the River Rhine. This research is concerned more with water quantity than with water quality. Support for the latter type of research is the task of the International Rhine Commission (IRC). The publication Monography of the Rhine Basin (CHR/KHR, 1976) contains maps of land use, soil type, geology and elevation covering the Rhine drainage basin. The scale of the maps is 1:1,500,000. Although more detailed maps are available, the advantage of using these maps was that they provided a consistent interpretation of the geographical variables in the entire basin. Apart from these geographical data, we expected that time series of monthly temperature and precipitation and discharge could be readily obtained. To obtain a digital geographical database, we digitised the land use, soil type and elevation maps. We undertook steps to collect the hydrological and meteorological data. We enjoyed much cooperation from the hydrologie institutes in supplying the discharge data. However, we found difficulties in collecting the meteorological data: only the Swiss meteorological database is easily accessible for scientific research. All data were stored in a raster GIS with a spatial resolution of approximately 10 km2. Using these data we built a prototype of the model which can be considered an integrated GIS-based water balance model for the River Rhine. The model algorithms are implemented in the GIS using generic functions and a set of tools for hydrologie and géomorphologic modelling (Van Deursen and Kwadijk, 1993). The model produces quite accurate results. Except for one station, the difference between modelled and measured annual discharge is less than 5 per cent, both for subcatchments (3,000 km2 and larger) and for the whole catchment. The model can also simulate accurately time series of monthly discharge in different sub-basins. Comparing model estimates of average areal monthly evapotranspiration with published data shows that the model is capable of representing the evaporation losses accurately both in the Alpine basins and in the lowland basins. Snow water storage is also quite well represented at different altitudinal zones in the Alps. Parallel to the development of the RHINEFLOW 1.0 model, research projects were carried out at the Universities of Utrecht and Amsterdam on the fluxes of sediment transport dynamics of the River Rhine. This research included soil erosion, suspended sediment transport and sediment deposition rates in the floodplains (Asselman, 1995; Middelkoop, 1995, Van der Drift, 1995). At the Netherlands Institute for Environmental

50

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Research (RIVM), a study was carried out on the diffuse sources of contaminants in the Rhine basin.

Development and data collection for RHINEFLOW version 2.0 As the RHINEFLOW 1.0 model showed acceptable results, the project was incorporated into a large research project initiated by the CHR/KHR Commission in 1991. The objective of this project is to investigate the impact of environmental change in the Rhine basin. This includes effects on different river functions such as navigation, industrial and public water supply and water supply for energy production. Part of this project also includes the development of an updated model version of the RHINEFLOW 1.0 model, referred to as RHINEFLOW 2.0. This update will include:

¡ improvement of the model by obtaining a higher spatial resolution (1 km2 instead of 10 ¡

km2) and a higher temporal resolution (10 days instead of 1 month); extension of the model with modules to simulate fluxes of sediment, nutrients and pollutants. For this purpose the results of the different studies mentioned above will be incorporated.

The main objective during the first project phase was to find out whether it was possible to develop a water balance model for the River Rhine and to investigate how reliable this model was in representing this water balance. The purpose of phase two was to improve the prototype model to an operational version that could be used over a longer time period by different organisations. To develop this upgraded model a great amount of data is necessary. The use of the model by the water management authorities also means that, in this phase, the data must be collected in a more formal way than during the first phase. During the first phase a limited data set, time series for temperature, precipitation and discharge between 1956 and 1980, allowed for the calibration and validation of the model. As described, the other data types such as maps representing land use and soil type were made usable by the model in a rather ad hoc manner. The updated model version, however, should allow for simulation of recent discharges, such as the ones causing the floods of 1993 and 1995. This requires a data collection strategy that guarantees continuous data flow to update the database. Such a strategy needs agreements between the users and the data providers that updated data series will be made available now and in the near future. These requirements relate not only to hydrologie and climatic time series, but also to information on water demand, land use and water and land management practices. When such practices change, the new model should be able to simulate the effect of the changes on the water balance. This means that these data should also be easily updated. The combination of geographical data to develop a consistent database is much more difficult since we cannot use the original maps published by the CHR/KHR (CHR/ KHR, 1976). For the second project phase, geographical data on a much higher resolution is necessary which means that these data come from different countries, are collected from maps that have various scales, use different reference levels and originate from databases that use different classifications. The strategy being followed is to collect the data through the CHR/KHR as well as to use data that had been collected during the previous study. There are two factors that facilitate the data flow during this phase:

A GIS FOR HYDROLOGICAL MODELLING OF THE RHINE

51

¡ the potential users, the institutes for water management, are involved in the project. ¡

They supply some of the data, or have good relations with other data providers, such as the meteorological surveys; the floodings of 1993 and 1995 raised funds for this type of research, which allowed the institutes for water management to spend time and money on these issues.

To obtain the geographical data we are following two parallel strategies. In the first place the CHR/KHR approaches the data providers and gathers the data requested by the researchers. Table 4.1 shows the data providers and the types of data used. Apart from data collection for this single research project, the CHR tries to make arrangements with the providers for the supply of future updates in the desired format. With this construction the CH R functions as an intermediary between the researchers and the data providers. The CHR does not intend to become a data provider itself: it intends more to be a subscriber to the data that it needs for its own research. As the CHR members are the national representatives for water management, many problems with respect to data protection can be avoided. For example, it is not easy for a researcher from the Netherlands to obtain a DTM from Switzerland because this information is protected. The Swiss Hydrological Survey, however, has access to this information. The GIS group of the CHR combines the data and converts them into maps covering the Rhine basin. Since data originate from several institutes in different countries, converting the database into one consistent database for the entire basin is not easy. The different countries use various systems to classify their land use and soil types and they have different reference levels for altitude measurements. Water level is referred to as a level in metres above sea level in the Netherlands and in Switzerland, while in Germany different gauges use different reference levels. With respect to the hydrological and meteorological data, they use different procedures to measure different variables. Because of these different procedures and interpretations, some rather obscure changes occur on the borders of different countries. Examples include the disappearance of 100 m3/s of water in the Rhine between Germany and the Netherlands; cliffs of several centimetres to several metres that form the border between countries, and winding boundaries between soil and land use types that run parallel to the national borders, even when both countries use the same classification. Some of these artefacts, such as the differences in altitude reference level, can be solved quite easily; others are more difficult to solve. For example, the European soil map forms the basis for soil type in the EU part of the basin. The digital version can be used directly. For Switzerland we use another database, the so-called Hektaren raster. This contains a land use classification. Attributed to this classification is an indication of the soil types that are preferred for different land uses. In this latter case the user must choose which soil type is attributed to each cell, taking into account other geographical information such as altitude. Another problem arises when a variable, such as land use, is attributed to administrative units. In that case the original data comprise polygons representing these units. Attributed to these polygons are the numbers of square kilometres covered by a certain land use type. The exact location within the polygon is not known. Since we use a raster-based GIS, this type of information cannot be converted directly into the model database. A last example of a classification that is difficult to compare between countries is one that gives semi-quantitative interpretations of the intensity of a process. For example, an area where soil erosion is severe should be interpreted in quite a different way in Germany than, for example, in Spain.

Table 4.1 Sources of geographical data used in the RHINEFLOW 2.0 model

A GIS FOR HYDROLOGICAL MODELLING OF THE RHINE

53

Hitherto these differences could be solved easily since the hydrological model used so far employed classifications for land use and soil type with less detail than the data provided. However, we foresee that a great deal of effort will have to be put into the creation of a more or less consistent database. Small differences in geographical variables will have a much larger effect on the results of models that describe sediment and pollutant fluxes through the basin than on the results of the original hydrological model.

Conclusion ‘What do you consider the largest map that would be really useful?’ ‘About six inches to the mile.’ ‘Only six inches!’ exclaimed Mein Herr. ‘We very soon got to six yards to the mile. Then we tried a hundred yards to the mile. And then came greatest idea of all! We actually made a map on the scale of a mile to the mile!’ ‘Have you used it much?’ I inquired. ‘It has never been spread out yet,’ said Mein Herr: ‘the farmers objected; they said it would cover the whole country and shut out the sunlight! So now we use the country itself as its own map.’ (Lewis Caroll, Sylvie and Bruno Concluded, borrowed from Van Deursen, 1995)

From what has been described, it is clear that the need for information may change even during one project. At the start of a project the researchers themselves sometimes do not know exactly what they will need, which often results in letters to data providers requesting all the data that they have on a particular item. Research questions also change in time. This requires different types of information, different spatial and temporal resolutions, new methods to combine data from different sources and so on. We believe that it is not realistic to think that a database can be built which will include all the information that is necessary to solve questions now and in future—apart from the database described by Lewis Caroll. Nor can we expect the data providers to gather all the kinds of information that the researchers say they need and to distribute this freely amongst them in the format they want. With respect to the brief description of the problems encountered with the use of different reference levels and classifications, we are also rather pessimistic about the possibility of early agreements on the classification and description of the environmental data. For example, a simple variable such as average daily temperature, measured since the seventeenth century, is determined by varying methods in different countries within the EU. As databases are designed for national or even regional purposes, the classifications used will also serve these local purposes. This may change if we all consider ourselves being more European than French, German, English or Dutch, but this change will take quite a while. One could even argue that it is better that different methods exist in parallel as no-one knows the best one. What we can expect from the data providers in the mean time, however, is that they advertise that they exist. At the start of each large-scale environmental research project, much time is spent locating the databases that are needed and which can be consulted. The names of many databases are often known, but it takes a while to determine what data are in the database and which procedures should be followed to get them out. In many cases, as in this project, the researcher is dependent on the willingness of the persons contacted.

54

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

From the researcher’s point of view, it would be an enormous advantage if the data providers could prepare a bulletin in which they make clear what kinds of data they own. This information includes units, classifications, georeferences, temporal and spatial resolution, methods used for collection, area covered, etc. How, as well as in what format the data can be obtained, should also be provided. Consequently we would suggest that a European framework for geographical data exchange in its most simple form could be an Internet site where data providers advertise their products. The advertisement should contain at least an indication of what kinds of data can be expected when the database is consulted and where further information about the database can be obtained (preferably again an Internet site). More detailed information on the data should be provided at this address. If a user believes he needs the data from a certain database, he can make his own agreements with the supplier on the conditions necessary for the use of the data. The creation of such a central data information site needs some agreements on the format of the advertisement and preferably some rules concerning the conditions on which a data supplier is allowed to advertise on the site. We are aware of the problem that there must be incentives for a data provider to make such advertisements. On the other hand, if non-commercial database managers are able to show that their data are valuable to many researchers (by counting the number of visits to their own Internet site) they have strengthened their arguments for funding their efforts for database maintenance.

Acknowledgements The development of a geographic information system for the River Rhine is carried out within the framework of the project Effects of Climate Change in the Rhine Basin of the International Commission of the Hydrology of the Rhine Basin (CHR/KHR). We thank Nathalie Asselman for her comments on the manuscript and Marcel de Wit for providing information on the addresses of database providers.

REFERENCES ASSELMAN, N.E.M. (1995). The impact of climate change on suspended sediment transport by the River Rhine, in ZWERVER, S., VAN ROMPAEY, R.S.A.R., KOK, M.T.J. and BERK, M.M. (Eds), Climate Change Research, Evaluation and Policy Implications, Studies in Environmental Science 65A, pp. 937–42. Amsterdam: Elsevier. CHR/KHR (1976). Le basin du Rhin, Monography Hydrologique/Das Rheingebiet, Hydrologische Monographie, Parts A, B and C. ‘S Gravenhage: Commission Internationale de l’hydrologie du basin du Rhin/ Internationale Kommission für die Hydrologie des Rheingebietes. KWADIJK, J.C.J. (1993). The impact of climate change on thé River Rhine, Netherlands Geographical Studies, 171. Utrecht: Geographical Institute, University of Utrecht. MIDDELKOOP, H. (1995). The impact of climate change on the sedimentation rates on the embanked floodplains in the Netherlands, in ZWERVER, S., VAN ROMPAEY, R.S.A.R., KOK, M.T.J. and BERK, M.M. (Eds), Climate Change Research, Evaluation and Policy Implications, Studies in Environmental Science 65A, pp. 931–7. Amsterdam: Elsevier. VAN DER DRIFT, J.W.M. (1995). The effect temperature change on soil structure stability, in ZWERVER, S., VAN ROMPAEY, R.S.A.R., KOK, M.T.J. and BERK, M.M. (Eds), Climate Change Research, Evaluation and Policy Implications, Studies in Environmental Science 65A, pp. 923– 30. Amsterdam: Elsevier.

A GIS FOR HYDROLOGICAL MODELLING OF THE RHINE

55

VAN DEURSEN, W.P.A. and KWADIJK, J.C.J. (1993). RHINEFLOW: an integrated GIS water balance model for the River Rhine, in KOVAR, K. and NACHTNEBEL, H.P. (Eds), Application of Geographic Information Systems in Hydrology and Water Resources Management. Proceedings of the Conference on HYDROGIS, 211, 507–19. Wallingford: IHAS. VAN DEURSEN, W.P.A. (1995). Geographical information systems and dynamic models, Netherlands Geographical Studies, 190. Utrecht: Geographical Institute, University of Utrecht.

CHAPTER FIVE

The MEDALUS georeferenced database: an application used for land degradation research in the European Mediterranean ERIK CAMMERAAT AND HEIN PRINSEN

Introduction One of the general foci of environmental research is currently the degradation of semi-arid and Mediterranean environments, as they are considered to be vulnerable and fragile ecosystems (Perez-Trejo, 1994; Imeson, 1995), which certainly are affected by socioeconomic driven perturbations (Leeuw, 1994, 1995) and possibly also by climatic change (Palutikof, 1994). In environmental research projects, such as in M E DALU S (Mediterranean Desertification and Land Use) large quantities of high quality data are gathered as these projects cover large areas and deal with complex interdisciplinary research questions (Brandt and Thornes, 1996). Working in large research teams also requires the possibility of exchange of large data sets between the research participants, and the dissemination of data into the scientific and public domains. The problem with most of the existing data sets is that they are usually difficult to access and are formatted in many different binary and non-binary forms. To be able to benefit fully from these resources, data need to be both accessible and transparent. It is obvious that data on environmental research should be stored or presented with their geographical reference (for example, connected to a map) or with their specific coordinates. Although many generalised studies have been undertaken to gain a general idea of some aspects of environmental degradation (e.g. CORINE, 1992), these impressive products are derived from various non-uniform data sources, are very thematic and generally lack a process-based nature. GIS applications are also used to evaluate environmental properties and qualities (Garg and Harrison, 1992; Theocharopoulos et al., 1995) The georeferenced database presented in this chapter was developed within an interdisciplinary research project studying various aspects of Mediterranean desertification and land use—MEDALUS (Brandt and Thornes, 1996). This research aims to understand relevant processes involved in land degradation and land use change as well as to provide some tools for the mitigation of land degradation and for land rehabilitation. The project has several interconnected themes: 57

58

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 5.1 Overview of the European Mediterranean showing the MEDALUS field locations except for those of Portugal.

¡ obtaining field-based information on land degradation derived from work at various scales, ranging from point observations to remote sensing images;

¡ modelling of land degradation processes by deterministic computer simulation models which include the importance of vegetation for erosion and runoff generation;

¡ study of climate change and the development of climate change scenarios; and ¡ study of socio-economic aspects controlling land use change. One of the important objectives of the MEDALUS project is to increase the understanding of the processes and origins of soil degradation and desertification in regions of the European Mediterranean which required data collection on these processes. The study was carried out at eight core field sites all over the European Mediterranean, and at three large demonstration areas (Figure 5.1). At these sites biotic and abiotic environmental processes were studied at different levels of detail on specially designed measurement sites (Cammeraat, 1996). At the start of the project, many essential data were found not to be available and, if they were, certainly were not comparable, owing to the utilisation of different measurement and sampling techniques. The MEDALUS project was in a more or less favourable position to generate a special database of the properties required. At the beginning of the project a harmonised sampling program was defined (Cammeraat, 1993,1996). The objectives of the sampling programme fitted into the development of a deterministic hillslope hydrology and vegetation growth model (Kirkby et al., 1993). It also made comparison possible between the eight selected and representative field sites. Another advantage of the system presented is that the program is relatively easy to use and data are stored transparently, which makes it also very suitable for training purposes.

THE MEDALUS GEOREFERENCED DATABASE

59

Not only could students be familiarised with aspects of environmental degradation but policy-makers could also gain a rapid impression of these problems, as they are generally far away from research.

Methods Obtaining the necessary data New data from f ield and laboratory observations and measurements As explained above, the project created the opportunity to compile a completely new data set derived from field measurements. To benefit fully from this position, protocols had to be developed which aimed at standardisation and harmonisation of the data collection and processing. At the outset of the project a plenary meeting was held with both the modellers and the field research teams. At this meeting the modellers explained their ideas of what the simulation models looked like and, more importantly, what types of data they thought would be needed as input. On the other hand, it was clear that the field teams would not be interested in data-gathering only. The first agreements on data and field locations were finalised in a protocol, the MEDALUS Field Manual (Cammeraat, 1993), which was revised and updated as the project progressed. The data were collected in the field and consisted of several types ranging from continuous time series to one-off measurements. Many abiotic and biotic parameters were measured, as indicated in Table 5.1. This was done at the eight core field sites and continued for more than three years; it is still being carried out at three sites. Detailed site descriptions can be found in Brandt and Thornes (1996). The selection of field sites at the eight core site locations was dependent on some minimum requirements: the sites should be representative and at least have two types of land use. In some cases existing measurement sites were used and extended, whereas in other places new sites were developed. The data were sampled following the harmonised field and laboratory procedures (Cammeraat, 1993) and individual researchers were trained in workshops to develop similar sampling techniques. The results obtained from field and laboratory measurements were filled in on predefined electronic spreadsheets (Lotus *.wkl format) to make standardised dataprocessing possible and these were sent to the central database administration where they were further processed. The data were stored in a temporary format on hard disk and were distributed by floppy disk on request, or could be retrieved directly from a UNIX server. Meteorological data from the weatherstation at the field locations were processed to ten minute or hourly intervals. At some places no weatherstations was set up, and buying data from the official authorities was quite expensive. Climatological data from official weatherstations, kept in the European climatological database at the University of East Anglia in Norwich, UK, were added to our database, covering at least ten years of observation on precipitation and temperature data.

Collection of external a7nd dispersed data Basic data such as topographical maps, aerial photographs and other published material were also collected but this was difficult as we had to deal with various authorities and publishers in many different countries. For each site location official permission was obtained

60

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 5.1 Parameters being measured in the MEDALUS project at eight field locations

a

Not measured at all sites Source: Cammeraat (1996)

for clearance of copyright for electronical reproduction by scanning of topographical maps and aerial photographs. Various problems were encountered as some authorities seemed to have no official policy at all and gave permission without any hesitation, sometimes even wondering why permission was asked, whereas others had a very well developed commercial policy like the French IGN. Access to, and permission from the Greek authorities was only possible by the kind help of our Greek colleagues. In some other cases it was not clear at all who

THE MEDALUS GEOREFERENCED DATABASE

61

owned the copyrights. It will be understood that all these inquiries and arrangements were highly time consuming. The geological maps were a separate matter, as they all had different layouts and legends and were often not suitable for scanning. Furthermore, only very small parts of the maps were necessary, which needed revised legends. These were simplified and digitised in a general format with topographical information included. This also avoided the problem of obtaining agreement from the publishers of these maps. All maps and aerial photographs were either scanned or digitised and were formatted into bitmap formats or vector-based maps. Advantages of a georeferenced database The data collected in the MEDALUS program were diverse in nature and format. The standardised sampling procedures also needed a standardised data retrieval system. The data that were collected at the eight localities from the European Mediterranean at different scales included point, plot, transect and catchment scale data. Therefore the exact geographical positions of the data had to be included in the database. Exact georeferencing was a critical condition if the database system were to be developed. Before selecting the actual database system it became clear that it should have several other properties:

¡ It should have an intuitive user interface and should be running on standard computers,

¡ ¡ ¡ ¡

allowing the program and its application to be learnt quickly. This means that it should be possible to use and apply the program within a day, without detailed technical training. It was found to be very important that the researchers should easily get access to the data of the other localities. A fancy database which needs several weeks of training was considered to be useless for this type of application. A flexible set-up of the basic database structure so that a layout change in the future would not need considerable reprogramming. Addition of other types of data, as well as new data should be possible without great effort. It should be able to handle different types of data such as text documents, tables, images, and raster and vector maps. The program should be robust and relatively low cost. Analysis of data should be possible.

When trying to find an actual system which fulfilled these necessities, it became clear that generic database programs, as available in 1993, could not handle these different types of data without considerable additional development of programs. Several interesting extended databases existed with related types of information. However, these databases would need a fair amount of restructuring if they were to be useful for the present purpose as they were principally set up for other applications. The possibility of a structure based on hypertext was also considered but this would have needed great programming effort. The Geomanagement System (GMS, 1994) was found to be a interesting intermediary between the data and the users, and after a first test application it was decided to put the data into a database with GMS as the interface. The concept of GMS The concept of geomanagement combines the multiple criteria approach with the technical possibilities of multiple windowing and simultaneous communication among different

Figure 5.2 Scheme showing the structure of the geomanagement system.

THE MEDALUS GEOREFERENCED DATABASE

63

Figure 5.3 Organisation of the GMS application.

sources of information to support integration and management of all kinds of data in a georeferenced data management system. By using the unique georeferences (latitude, longitude coordinates) the geomanagement software supplies a powerful tool to exchange, integrate, analyse and report on the various dimensions of a problem across many geographic locations. For example, information such as thematical and topographic maps, GIS files, remote sensing images, text and spreadsheet files and pictures can be interactively compared by linking them by their geographical reference. Because GMS allows simultaneous communication among different software applications, it is no longer necessary to convert all data to one unique format. This, together with the use of multiple windowing, makes it easier to exchange, analyse and report on the various data from different geographical sites (Figure 5.2). A further advantage is that any existing map, field data or aerial photograph can be added to the system without problems like rescaling or complex processing. Figure 5.3 shows the layout of a typical GMS application. In an atlas a number of georeferenced maps are stored which can be used either as a background to locate information resources (for example, positions of research sites and locations of measured data) or as information maps that can be compared and give information on ‘what is where’ relations. Documents or data files are stored in a thematic catalogue by attaching them

64

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 5.4 Contents of a typical card showing information on data and field sites.

to a card on which the georeference and some general description of the attached document(s) is specified (Figure 5.4). Each card is linked as an icon marker to the atlas maps by the georeference (latitude, longitude coordinates) of the document. The cards can be selected in different ways, using thematic keyword indexes, geographical references, dates, numerical selection criteria and so on.

Results and discussion The application on CD-ROM In addition to the data obtained from new measurements and the gathered base data such as topographic maps, other information was also stored on a CD-ROM for data exchange (Cammeraat and Prinsen, 1994) using the GMS system as a user interface or shell around the data. The project reports of the participating research groups were also included, as well as the bibliography of the field sites. In the editing phase of the georeferenced database all data files were screened, edited and georeferenced and attached to the GMS system. An important inclusion was the MEDALUS hillslope model (Kirkby et al., 1993) which was specially designed for modelling hillslope hydrological processes and vegetation growth under Mediterranean conditions. Processed monthly, METEOSAT images over 1993 from the western and eastern Mediterranean were also included, which were processed and obtained from EARS in Delft. Two versions of the CD-ROM were available, both with the same data but with different versions of the GMS program: a runtime version at low price, and a second version with the full GMS program enabling the extension of the georeferenced database with new data. In total, 35 copies were produced for internal use within the project. Further

THE MEDALUS GEOREFERENCED DATABASE

65

dissemination of this CD-ROM is in progress. In the current phase of the MEDALUS project, the georeferenced database will be extended with new results, including results and data from the demonstration areas. The possibility is currently being explored of making the database accessible via the Internet. The database index can be retrieved from the world wide web but the actual data themselves must be retrieved from the research groups that obtained them in the field as there is no general consensus on the reproduction rights of the actual data. The georeferenced database is increasingly being used as a data source for the joint publication of research papers, comparing research results from the different research sites, as well as for modelling purposes.

Application of data for modelling using climatic and socio-economic scenarios The data in the georeferenced database were originally sampled in conjunction with the development and validation of the MEDALUS catena hillslope model (Kirkby et al., 1993). The data can be used to simulate the hillslope runoff, erosion, soil moisture and vegetation response of the field sites. After setting up the model for the specific field site conditions, measured meteorological data can be used to run the model. The outcome of the simulation runs can be used to validate the model or to study specific environmental problems. One could also apply certain climatic scenarios, derived from GCM results and other climatological research (Palutikof et al., 1996). Possible effects of drought or changes in seasonal effects of precipitation on the geo-ecosystem can be evaluated with the model. A similar procedure can be followed to implement the effects of socio-economic changes on the environment, for instance resulting from land use changes.

Case study using the application for a comparative degradation analysis of the El Ardal (Spain) and Petralona (Greece) sites In this section an example is given of an application of the geomanagement system. Two sites, one in south-east Spain—El Ardal, Murcia (Lopez-Bermudez et al., 1996)—and another in northern Greece—Petralona, Thessaloniki (Diamantopoulos et al., 1996)—are studied with respect to their soil types, vegetation cover and erodibility. The two areas are interesting as they show many similarities from a lithological, physiographical and current land use point of view. However, their land use histories are very different and this factor may explain why these two areas show such a different behaviour with respect to erosion and runoff (Cammeraat et al., in prep.). By combining data from both sites an analysis can be performed and facts can be retrieved. The first step is to select the Murcia region as a search zone on the map of southeast Spain. A first impression of the site can be obtained from its landscape by opening photographs on the screen (Figure 5.5). Then a query is carried out on the topics needed, for instance soil profiles, aggregate stability, infiltration, runoff and sediment yield as well as vegetation cover characteristics. This gives access to several data files containing the requested information. Related information is also indicated on maps, such as maps of soil types, photographs of some of the soil profiles and the chemical and textural composition of the soil itself (Figure 5.6). A map presenting the field site layout can be opened and kept in a window for spatial reference, as the catenal position determines many properties of the soil and its surface. A similar procedure can be followed for the Greek site. Data can easily be extracted from the retrieved documents to build new documents containing the

66

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 5.5 Overview of the landscape of two sites of the MEDALUS georeferenced database.

Figure 5.6 GMS application with soil map, soil profile description and pictures of the soil profile described.

THE MEDALUS GEOREFERENCED DATABASE

67

Table 5.2 Some general characteristics of the two field sites

data needed by copying and pasting. Tabulated data, like the aggregation and infiltration data, can be incorporated into a new table or graph to make intersite comparison possible. The new documents and tables can form the backbone of a paper or study on specific aspects of these two sites. From this analysis we can combine several data (Table 5.2 gives some of these data), which are important for the analysis. The striking difference between the two sites is that runoff generation and erosion on both sites are of different orders of importance. This is also shown by the results of rainfall simulations carried out at both sites. Furthermore, soil profile development and thickness are different and generally more soil is present at the Greek site. This difference is partly due to different types of limestone and their weathering products, and partly to the degree to which these soils have undergone soil erosion. Both sites are currently grazed by sheep and goats, but their grazing history is different. At the Spanish site there is a history of grazing over several millennia, whereas grazing intensity in this part of Greece has increased considerably since the 1920s when it was colonised by Greeks expelled from Turkey. It is therefore suggested that this difference in the length of grazing pressure may also be reflected in the properties of the soil. It is well known that bare limestone with only an A horizon on top has generally a very high infiltration rate and produces little runoff. This is the case in the Spanish site. On the Greek site, however, there is still a clear B horizon present, which is at the moment strongly degraded by erosion and trampling processes, at some places even resulting in gully initiation. This hypothesis is currently being further tested and analysed (Cammeraat et al., in prep.). Another example of an application is the possibility of map overlaying. Sections of one map can easily be marked, by defining their surroundings, and projected on other maps as a kind of overlay. However, this possibility does not take into account the distortions due to different projections of maps. This might be a serious problem especially with respect to aerial photography. Improvements to be made The contents of the database itself can be improved by a more balanced content, as some sites provided only the minimum data set whereas others provided information in much more depth. The addition of processed Landsat and Spot remote sensing images would also increase the value of the data set. Furthermore, a much wider application of georeferenced maps and photographs could be used, providing even more detailed spatial information.

68

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

It is not possible to program a query sequence and to define a combined output format, as is possible in many databases running under query languages such as SQL. This may be seen as a disadvantage, especially in the case of event-based data and other time series. No query can be carried out within the data file itself, but only on the attached parameters on the card of the data file (document). From the technical point of view, several improvements are proposed. One of the utilities to be added is the possibility of querying within documents. To increase the use of map overlay possibilities, a transformation utility would be helpful to overcome the problem of map distortions due to different projections. The number of open applications in a GMS session can also be rather large and could be reduced by opening multiple documents in one application. At present all documents run in a separate window with their application, even if these documents are of the same type. This improvement would increase the speed of the computer system as Windows applications generally use a lot of memory. By improving these points an even more powerful tool will be available, especially when it is applied to the analysis of data. Another, Windows-related problem is that the graphical output is dependent on the resolution selected by the user and the resolution of the graphic images stored on the CDROM. When these are not compatible, poor quality images can result.

Conclusion A georeferenced database management system (DBMS) was developed to store harmonised data from eight field sites in the European Mediterranean. At these sites research was carried out on the processes involved in land use change and land degradation. The use of georeferenced databases makes it possible to compare real data in a very flexible and transparent way. It also allows different types of data to be compared and connected via their georeference. The system enables the analysis of environmental data and it allows data from various sources to be combined. Its simple user interface lowers any resistance to working with large spatially organised data sets. Under normal database applications this is often a problem, as many environmental scientists have a limited experience with computing in general and database applications in particular. These qualities of the application make it suitable too for students, managers and policy-makers, enabling them to work with real data, which could improve their understanding of actual environmental problems. By combining the georeferenced database with computer simulation models it is possible to work with field data from different sites and to estimate the evolution of the sites under different scenarios of disturbance. An example is given where the georeferenced database is being used for the evaluation of land degradation at two Mediterranean sites that were studied in the MEDALUS project.

Acknowledgements The work for this paper was carried out as a part of the MEDALUS project. MEDALUS is funded by the European Commission under the European Programme on Climate and Natural Hazards (EPOCH) EPOC-CT90–0014-(SMA) and EV5V 0128, the support being gratefully acknowledged. The cooperation of all the teams involved in the project has made the database what it is now and is greatly appreciated.

THE MEDALUS GEOREFERENCED DATABASE

69

REFERENCES BRANDT, J. and THORNES, J.B. (Eds) (1996). Mediterranean Desertification and Land Use. Chichester: Wiley and Sons. CAMMERAAT, L.H. (1993). MEDALUS Field Manual, v3.1, with contributions from Cammeraat, L.H., CLARK, S.C., COLLIN, J.J., IMESON, A.C. and MELRA, J.BRISTOL: University of Bristol. CAMMERAAT, L.H. (1996). The MEDALUS core field programme: An overview of sites and methodology, in BRANDT, J. and THORNES, J.B. (Eds), Mediterranean Desertification and Land Use. Chichester: Wiley and Sons. CAMMERAAT, L.H. and PRINSEN, H. (Eds) (1994). The MEDALUS Application on CD-ROM. Chaumont-Gistoux and Amsterdam: Da Vinci/FGBL/University of Amsterdam. CAMMERAAT, L.H., DIAMANTOPOULOS, J. and ROMERO-DIAZ, M. (in prep.). Soil erosion on hillslopes on limestone in Halkidiki (N.Greece) and Murcia (SE Spain): its implications for landscape vulnerability. DIAMANTOPOULOS, J., PANTIS, J.P., SGARDELIS, S.P., IATROU, G.D., PIRINTOS, S., PAPATHEODOROU, E., DALAKA, A., STAMOU, G., CAMMERAAT, L.H. and KOSMAS, C. (1996). The Petralona and Hortiatis field sites, in BRANDT, J. and THORNES, J.B. (Eds), Mediterranean Desertification and Land Use. Chichester: Wiley and Sons. CORINE (1992). CORINE Soil Erosion Risk and Important Land Resources in the Southern Regions of the European Community. Luxembourg: Office for Official Publications of the European Communities. GARG, P.K. and HARRISON, A.R. (1992). Land degradation and erosion risk analysis in S.E. Spain: A geographic information system approach, Catena, 19, 411–25. GMS (1994). GMS-DECIDE for Windows Manual, vl.2. Chaumont-Gistoux, Belgium: Da Vinci Consulting. IMESON, A.C. (1995). The physical, chemical and biological degradation of the soil, in FANTECHI, R., PETER, D., BALABANIS, P. and RUBIO, J.L. (Eds), Desertification in a European Context: Physical and Socio-economic Aspects, pp. 153–68. Luxembourg: Office for Official Publications of the European Communities. KIRKBY, M.J., BAIRD, A.J., LOCKWOOD, J.G., MCMAHON, M.D., MITCHELL, P.J., SHAO, J., SHEEHY, J.E., THORNES, J.B. and WOODWARD, F.I. (1993). The MEDALUS slope catena model: a physically based process model for hydrology, ecology and land degradation interactions, MEDALUS I Final Report, pp. 97–139, Bristol and London: University of Bristol/ King’s College. LEEUW, S. VAN DER (1994). Social and natural aspects of degradation, in TROEN, I. (Ed.), Global Change: Climate Change and Climate Change Impacts. Focusing of European Research, Proceedings of the symposium held in Copenhagen, Denmark, 6–10 September 1993. Brussels: European Commission. LE EUW, S. VAN DER (1995). L’homme et la degradation de l’environnement. Actes des XVe rencontresd’archelogie et d’histoire d’Antibes, p. 514. Sophia Antipolis: Editions APDCA. LOPEZ-BERMUDEZ, F., ROMERO-DIAZ, A., MARTINEZ-FERNANDEZ, J., MARTINEZFERNANDEZ, J. (1996). Soil and vegetation cover, in BRANDT, J. and THORNES, J.B. (Eds), Mediterranean Desertification and Land Use. Chichester: Wiley and Sons. PALUTIKOF, J. (1994). Mediterranean land use and desertification—the MEDALUS project, in Troen, I. (Ed.), Global Change: Climate Change and Climate Change Impacts. Focusing of European Research, Proceedings of the symposium held in Copenhagen, Denmark, 6–10 September 1993, pp. 165– 72. Brussels: European Commission. PALUTIKOV, J.P., CONTE, M., CASIMIRO MENDES, J., GODDESS, C.M. and ESPIRITO SANTO, F. (1996). Climate and climate change, in BRANDT, J. and THORNES, J.B. (Eds) Mediterranean Desertification and Land Use. Chichester: Wiley and Sons. PEREZ-TREJO, F. (1994). Desertification and Land Degradation in the European Mediterranean, Report EUR 14850 En. Luxembourg: European Commission. THEOCHAROPOULOS, S.P., DAVIDSON, D.A., MCARTHUR, J.N., TSOULOUCHA, F. (1995). GIS as an aid to soil surveys and land evaluation in Greece, Journal of Soil and Water Conservation, 50(2), 118–24.

CHAPTER SIX

Developments in cross-border standards for geodemographic segmentation RICHARD WEBBER

Introduction Most commercial GIS applications are designed to facilitate the recording, storage and retrieval of geographic data. Many of these applications support operational activities, such as the management of resources or the control of distribution networks and involve the automation of manual, paper-based systems. A rapidly growing subdiscipline within commercial GIS is geodemographic analysis and targeting, an application area used by organisations that wish to target the marketing of their products to discrete types of consumer. Unlike other applications of GIS, geodemographic applications inform strategic decisions governing product development, marketing strategy and distribution planning. This chapter addresses the emergence of geodemographics as a marketing tool and the development of common transnational classification codes for the implementation of micro-marketing strategies on a global basis.

Business context The business context in which geodemographic analysis and targeting has developed is generally referred to as micro-marketing. Whereas in the early postwar period marketing energies were focused on the promotion of major household brands using mass media (such as newspapers and television) to a mass and relatively undifferentiated audience, marketers have more recently sought to develop more differentiated products that will appeal to specific niche audiences and which can be promoted more cost effectively through media such as direct mail and door-to-door in order to reach a selected audience. In the era of mass marketing it was generally sufficient to describe target audiences in terms of general demographic categories such as age, income and social grade. Using conventional market research techniques, the advertiser could readily identify which categories were stronger buyers of his/her product and which newspapers and magazines were the most effective advertising media for reaching them. 71

72

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

As the emphasis of marketing shifted towards micro-marketing, limitations in the usefulness of conventional demographic measurements became apparent. They could not be used to analyse the behaviour of existing clients other than via expensive interviews; they could not be used to select from a rentable list the best prospects for a particular product; they could not be used to evaluate the profile of a retail catchment area or to quantify the market potential within localised markets that might be targeted using local press or door-to-door distribution. The emergence of geodemographic analysis and targeting is largely a response to the requirement of large advertisers for greater precision in the targeting of advertising and promotional budgets.

Using neighbourhood as a basis for market segmentation Geodemographics involves the classification of residential areas into distinct neighbourhood types based on statistical information about the consumers who live there. These classifications are typically built using very fine levels of geographic granularity (in the UK, Belgium and Netherlands, for example, units consist on average of only seventeen addresses) which, using address recognition systems, can be linked to the universe of national addresses. The link between the classification system and the universe of addresses means that any customer transaction record that includes the customer address can be used to define the types of neighbourhood in which users of specific products and brands are most over-represented compared with the nation as a whole. Marketing effort can then target such areas using a variety of media such as direct mail, door-to-door distribution, local press and retail outlets that can be selected on a geographical basis. Compared with other GIS activities, geodemographics typically involves planning and analytical rather than operational applications. Applications do not require a particularly high level of precision for locating geographical data but do require the ability to integrate GIS with statistical analyses, trade area definitions, the results of market research and systems for local marketing modelling. Today, most large retailers, financial services organisations, mail order operators, utilities and insurance companies will satisfy their needs for geodemographic analysis by using a proprietary geodemographic classification system. Often, this classification system will be accessed through proprietary geodemographic analysis systems, incorporating bespoke software, a variety of databases, configurations, and technical and applications training. Depending on the size and diversity of the national market, such systems typically identify between 62 categories (in the case of Claritas’ US classification system) and 28 (in the case of Experian’s classification system for the Republic of Ireland). For ease of use, these ‘fine’ classes are often arranged into a higher-order ‘coarse’ classification of between 8 and 12 groups. Table 6.1 shows the results of an analysis of Volvo owners by the twelve ‘coarse’ classes belonging to Experian’s UK classification MOSAIC, using respondents to the British Market Research Bureau’s Target Group Index Survey. The market segment richest in Volvo owners is labelled ‘High income families’ and has an average penetration rate 9 times that of the poorest performing group. Figure 6.1 shows photographs of the five fine classes within the high income families group. The photographs and nicknames are designed to communicate the distinctiveness

CROSS-BORDER STANDARDS FOR GEODEMOGRAPHIC SEGMENTATION

73

Table 6.1 Great Britain’s twelve MOSAIC lifestyle groups ranked by the penetration of Volvo cars using the Target Group Index market research survey

of each type in terms of values, attitudes and aspirations as well as base demographic characteristics. Multimedia is increasingly being used as a tool for explaining the classification concept to potential users and for disseminating information about target groups to customer facing staff such as store managers, salespeople and calls centre operators. Plate 3 shows an example of a multimedia screen used for explaining a geodemographic system to call centre staff.

Figure 6.1 Illustrations of the f ive f ine classes within the high income families group (a) clever capitalists.

74

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 6.1(b) Five classes: rising materialists.

Figure 6.1 (c) Five classes: corporate careerists.

CROSS-BORDER STANDARDS FOR GEODEMOGRAPHIC SEGMENTATION

Figure 6.1 (d) Five classes: ageing professionals

Figure 6.1 (e) Five classes: small-time Business.

75

76

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Geodemographics around the world Geodemographic classifications systems are now available in the USA and Canada, in eleven European countries, in Japan, Australia and New Zealand and in South Africa. Given the absence of a uniform statistical database and a common small area zoning system within Europe, let alone outside it, these classifications have been created on an ad hoc and opportunistic basis to meet local commercial requirements. Claritas and Equifax are the two most successful developers of geodemographic systems in the USA. CACI has developed systems in the USA, UK and Finland. Experian is the leading supplier of systems in Europe, South Africa and Australia, working collaboratively in a number of European markets with direct marketing organisations that have specialist access to local statistical databases. Bertelsmann (in Germany and Spain) has also established successful geodemographic classification systems. Worldwide, it is estimated that some 1,200 people are currently employed by vendors of geodemographic targeting systems and software. Two-thirds of these are employed in North America.

Data issues Table 6.2 shows the diversity of data sources that have been used by Experian to build MOSAIC classifications in the various countries covered by the MOSAIC international network. Each classification is built using a form of cluster analysis. UK MOSAIC, like its UK competitors, relies strongly on statistics from the 1991 national census as the criteria on which to base measures of similarity during the clustering process. However, because the units being assigned classification codes are individual postcodes, of which there are ten times as many as there are census enumeration districts, the census is supplemented by a variety of other data sources accumulated to postcode level. These include the UK electoral register, various measurements of consumer credit demand and bad debt, the local incidence of company directors, retail accessibility and the UK postal address file. Elsewhere in Europe different data sources are used. In the Netherlands, for example, where the public are wary of completing census forms, data are accumulated from market research surveys and from the databases of the country’s leading mail order operator. In Germany, where census statistics are available only at the suburb level, census statistics are reinforced with the results of visual assessment of individual buildings. France has good census statistics but these are no longer accessible for geographical units of fewer than 5,000 persons; Italy and Norway provide reliable and detailed census statistics; Spain has a street-level electoral roll containing age and education as well as mail order and car data. The Swedish MOSAIC operates at postnumber level using data from various personal and dwelling registers. Belgium has a mass of compiled statistics at street level. Besides the differences among markets in the sources of data that can be used to build geodemographic classifications, there are also wide differences in the level of granularity at which this data can be supplied. In the UK, Netherlands, Belgium and Ireland the units of classification contain an average some fifteen to twenty addresses. In Sweden and Spain the units are very much coarser. In the UK, Netherlands and Sweden the units that are classified are the most detailed levels of the national postcode. Elsewhere, because the units being classified lie at a lower level of resolution than the postcode, it is necessary to develop address recognition systems

Table 6.2 Data sources used in the MOSAIC classification in selected countries

Table 6.2 (continued)

80

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

for allocating the correct geodemographic type to an individual customer address record or a respondent to a market research survey. Given the variety of sources of information used to build these systems, their varying quality and indeed, in some cases, differences in their levels of geographic detail, a number of interesting modifications to standard clustering practice have been employed. For example, in MOSAIC, different weights have been given to different input variables to the cluster program and these weights have been altered dynamically according to the statistical reliability of the variable and according to the population size of individual zones. In both Experian’s and CACI’s systems, a separate cluster ran has been undertaken to provide better definition within particular parts of the country: in the case of CACI, London and Scotland, in the case of Experian the rural clusters. Most classifications are built using an interactive relocation based on a minimum sum of squares criterion. The consolidation of the final cluster solution in a set of coarse classes is often based on a stepwise fusion algorithm.

International compatibility When the various geodemographic classifications were first built, no attempt was made to constrain them within any cross-national classing system. Each system was built with the objective of maximising discrimination within that particular national market. The only international consistency was in the way that categories were imaged and in how the classifications could be accessed using PC software. During the late 1980s, when the US, UK, French, Italian, Dutch and German systems were being conceived, it was anticipated that clients would be interested in individual markets only. However, during the 1990s we have experienced a rapid growth in interest in global support systems for micro-marketing. If the global village appears to demand global brands, developed to meet the requirements for clearly defined demographic and psychographic target groups, then manufacturers, retailers and providers of financial services need to be able to target specific niches on a worldwide basis. The past five years has seen a significant increase in the number of US retailers and mail order operators launching their services into new national markets whilst Marks & Spencer and Ikea are examples of British and Swedish retailers that are now operating on a European rather than a domestic basis. Financial services organisations and retailers operating on a global basis seek organisations that can support them consistently in each of the markets within which they operate. With the growth of multinational target marketing, it is becoming increasingly clear that the 708 unrelated target markets that MOSAIC describes in ten different languages is hardly a satisfactory basis for targeting a consistently defined audience throughout the world. The question now is whether the supply of, in essence, local services through an international network was the appropriate way to meet the needs of multinational clients.

Towards a common international standard for geodemographic segmentation Most people assume that it is impossible to find demographically consistent types of neighbourhood on an international basis. Although there is no disputing the extent to which there exist in all countries clearly different types of residential area, the factors

Plate 1 The Baltic Sea region as perceived by the environmental (BDBP Project) and the hydrometeorological (BALTEX) communities.

Plate 2 The Baltic region according to the VASAB 2010 project.

Plate 3 Multimedia screen used to explain a geodemographic system to calls centre staff.

Plate 4 The distribution of EuroMOSAIC types in Munich.

CROSS-BORDER STANDARDS FOR GEODEMOGRAPHIC SEGMENTATION

81

which cause them to differ, it is often presumed, are so closely intertwined with the history of urban development, with climate and with national and local politics that no clear set of categories would exist at an international level. Even if it were the case, it might be argued, how could common categories be identified given the very real differences between countries in the sources of information and the level of geographic resolution at which it can be accessed. The more that one inspects the descriptions, photographs and textual interpretation given to the MOSAIC categories developed in each national market, the more one is tempted to challenge such assumptions. Far from viewing the statistical composition of areas as random profiles that could, with varying degrees of accuracy, be approximated to an average centroid, it seemed more productive to view these statistical profiles as manifestations of the original residential function of a neighbourhood in the historical evolution of a national economy. For example, whilst a particular cluster in a particular country could be characterised as having a high incidence of low income, older residents living in old, low rise, owner-occupied houses, with high levels of unemployment, few service workers and low levels of female employment, it could as easily be seen as an example of ‘Deindustrial legacy’, a type of declining community traditionally reliant on male jobs in coal, steel and other heavy industries which is too distant from major growth poles in the quaternary sector to sustain employment growth. Such types occur in many different locations: West Virginia, the South Wales coalfield, the Ruhr, the steel towns of northern Spain, around St Etienne in France and Maastricht in Holland. However different such areas may be in terms of architectural styles, language or local political autonomy, each area is characterised by a strong belief in collective action, by a traditional division of tasks between men and women, by reliance on co-ops and other forms of mutual organisation, by resistance to new and, in particular, foreign ideas, and by a preference for traditional food and drink. In such areas people are careful with what little money they have, they are not attracted to lifestyle symbols and do not aspire to foreign cars or foreign holidays. Other examples of genuinely international residential types include ‘articulate metropolitans’, people who are well educated, delay family formation as long as they can, are typically employed in the media and the influencing professions, whose social attitudes are liberal, whose orientation is international. Areas of ‘Dynamic families’ and ‘High rise social housing’ also occur in every market, albeit in different percentage distributions. Notwithstanding differences in the data available in each country, it would appear that patterns of social structures and urban development have led to common functional specialisation among residential neighbourhoods, even though their geographical locations within a city and their architectural styles may vary widely.

EuroMOSAIC: A Europe-wide standard for geodemographic segmentation Table 6.3 shows a common international standard of ten residential neighbourhood types—EuroMOSAIC—that has been adopted by the European members of Experian’s MOSAIC International Network for use in international micro-marketing activities. These ten types have been built up from the more detailed categories independently created for each local market. The purpose of the EuroMOSAIC classification is to provide multinational micromarketers with a consistent segmentation system that can be used on a Europe-wide basis. This can be used in a number of different ways:

Table 6.3 Distribution of EuroMOSAIC’s ten residential neighbourhood types in selected countries

CROSS-BORDER STANDARDS FOR GEODEMOGRAPHIC SEGMENTATION

83

¡ for defining target groups on a consistent basis; ¡ for ranking local markets throughout Europe on the basis of a consistent objective target group;

¡ for determining, on the basis of the results of user surveys in one national market, the ¡

most appropriate locations for focusing marketing activity when a product is launched in a new market; for comparing the user base of the same product in different markets so as to assess the appropriateness of a common European advertising strategy.

In order to build EuroMOSAIC from the MOSAIC’S of individual countries, Experian first identified what were felt to be key criteria for differentiating neighbourhoods. These were as follows:

¡ ¡ ¡ ¡

Age: Income: Urbanisation: Family type:

Young, middle aged, old High, middle, low Metropolitan, suburban, urban, rural Married couples, singles and childless couples, pensioners

Placing the individual clusters on the basis of their position on each of these four axes produces the possibility of 108 different combinations. National clusters were first assigned to the most appropriate of these 108 potential cells. The ten most frequently occurring cells were then allowed to form the nucleus of each EuroMOSAIC and the remaining clusters were then reassigned to whichever of the ten EuroMOSAIC types they then most closely resembled. Each of the ten types is given a standard colour, so that in any national map the colour red denotes ‘High rise social housing’ just as on a geological map it always denotes granite. Plate 4 shows an example of a EuroMOSAIC map in Munich. With the inclusion within the EuroMOSAIC classification of Coref’s French ‘ilotype’ classification in February 1996 and Seat’s Italian ‘cluster’ classification in January 1996, EuroMOSAIC has at last become a viable segmentation system for the large multinationals for whom the concept was originally designed. To test EuroMOSAIC we are now in the process of appending EuroMOSAIC codes to the results of market research surveys undertaken in different countries. Where questions have been asked on a consistent basis across different markets we are able to compare the profiles of distributions of customers or product users by MOSAIC in different markets. Early results show profiles to be most similar for consumer durables (e.g. cars, dishwashers) which have been introduced to the market this century. The level of consistency of product profiles is much lower in food and drink, markets which are driven by local and regional as well as national differences. Brand preferences (e.g. Renault, Sony) are less consistent than those of the categories of product to which they belong. Leisure activities, by contrast, show comparatively uniform profiles across different countries. Although neighbourhood classifications have been exploited most vigorously by consumer marketers, it should be remembered that they were originally developed to meet public sector applications. They have consistently shown themselves to be effective discriminators of demand for public services and of health and crime patterns. EuroMOSAIC has significant applications potential in comparing the incidence of target groups for public sector programmes across different European countries, in providing matching sample areas for particular pilot projects, and for identifying on the basis of pilot projects the localities across the continent which are most suited to the roll-out of specific public sector programmes.

84

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Conclusion EuroMOSAIC represents the ideal of being able to pull together data sets which measure along the same parameters and hence classify households in the same way across national borders. However, there is still significant scope for harmonisation of European data sources to facilitate greater integration and utilisation within geodemographic analysis. This should remain an important aim despite the emergence of external credit and life-style data. Such data, while powerful in building and enhancing national classifications, will always prove more difficult to harmonise across countries than the standard measures of housing, age, social grade and income currently available.

SUGGESTED READING MITCHELL, V.W. and McGoLDRICK, P.J. (1994). The role of geodemographics in segmenting and targetting consumer markets: a delphi study, European Journal of Marketing, 30, 289–333. TONKS, D.G. (1990). Pinning down geodemographics, Marketing Intelligence and Planning, 8(2), 1229–41. WATTS, P. (1994). European geodemographics on the up, GIS Europe, 3(4), 28–30. WEBBER, R.J. (1971). Liverpool Social Area Study, PRAG Technical Papers TP14. London: Centre for Environmental Studies. WEBBER, R.J. (1978). Parliamentary Constituencies: A Socio-economic Classification, Occasional Paper 13. London: Centre for Envrionmental Studies.

PART THREE

Multinational Database Products

CHAPTER SEVEN

European geoinformation data publishing: understanding the commercial industry VANESSA LAWRENCE

Introduction Organisations are run by data; they collect them, analyse them and, from these analyses, determine how they might be run more effectively. In the 1980s, as more organisations adopted geographic information systems, it became known that the cost of the hardware and software needed for a GIS often represented only approximately 30 per cent of the final costs of implementing the GIS project; the majority of the budget was being spent on data collection and conversion. These percentages are changing as the commercial data market starts to operate efficiently in the GIS arena. The 70:30 ratio is now only representative of high value, low volume situations involving start-ups of operational systems such as utilities. It is estimated that between 60 and 80 per cent of all data held by government departments can be classed as geospatial, where this is defined as any data that have associated with them some geographical referencing, including referencing to a national grid, a postcode system, latitude/longtitude or defined areas such as parliamentary constituencies. Geospatial information is often defined as having a spectrum ranging from:

¡ highly detailed topographical information where the geospatial element is fundamental to the nature and value of the information, such as maps; through to

¡ data such as national and regional macroeconomic statistics, company annual reports or information on parliamentary business where the extent of geographical disaggregation is usually minimal. As government and private sector organisations realised the huge investment they were making in the collection of data, the organisations’ accountants started to assess the advantages and disadvantages of allowing others to use the data that had been collected by their organisations as against using data created by third parties. This chapter explores the emergence of the commercial geoinformation data publishing market and considers its future.

87

88

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Data publishing: what is it? Geoinformation data publishing is part of the information publishing business. It can be denned generally as the process of making geoinformation data available to a wider audience in a format that is suitable for the end user. The data may be available in a proprietary file format such as DXF, MIF/MID, DON or SHP and can be supplied either on floppy disk, CD-ROM or over the Internet. No doubt the volume of data shipped on floppy disk will reduce rapidly as the use of the Internet for file transfer operations increases. Geoinformation data publishing can encompass activities such as on-demand publishing, where a map is produced with perhaps a house centred on the map which can either be sent to the customer in paper or digital form. Equally, the term applies to data sent by the Internet or to packaged volume products produced by several European companies which can be bought off the shelf and are ‘plug-and-play’ technology. An example of the ondemand system is Superplan operated by the Ordnance Survey in Great Britain, which is a system operated through third-party resellers, often bookshops, that have an up-to-date copy of the Ordnance Survey digital database which can be accessed and downloaded to the customer’s specification. An example of a packaged volume product would be RegioMap which is a pan-European statistical atlas produced on CD-ROM, copublished by Geoinformation International and EUROSTAT. As the European GIS industry continues to grow into non-traditional markets and systems become cheaper to purchase and easier to use, the commodity which will constrain the industry will be the lack of data available to meet the needs of the customer. No longer will organisations be prepared to spend huge sums on data collection or allow the time for its collection; they will certainly wish to be able to purchase standard data sets in an accessible form off-the-shelf and onto which they can overlay their own data. The emerging commercial data industry is trying to meet the needs of these customers but the difficulties of creating data products are greater than were at first envisaged. This chapter sets out the process, concentrating on the European data publishing market. The situation in North America differs from that in Europe as the price of raw data there is negligible. In the USA, as the raw data lack value, value is added at each stage of the process of creating a US-based product. The result is that there is now a plethora of products priced at $50-$100 for the US market and supplied through high street computer shops and mail order catalogues.

Creating data products: the importance of the market assessment Market research on the potential use, needed functionality and data fields in a new data product is difficult to undertake, but it can be very useful in shaping an idea for a final product. It can be difficult for potential users to envisage how they could use the data product if it were available—until they are introduced to a working product. Carefully designed questionnaires can be used about current data usage, but data products are often being planned as ‘future-proofed’ products so they tend to consider that individuals will be working slightly differently with technology by the time the product is available. Decisions must be made about the potential market for the geoinformation data product before the software front-end of the product is designed, the intellectual property obtained and the packaging or branding is decided. The following must be investigated and best guesses made:

EUROPEAN GEOINFORMATION PUBLISHING

¡ ¡ ¡ ¡ ¡ ¡

89

the market for the product; the size of the target market; how the target market can be reached and at what unit cost; the market’s perceived price for the product; the images to which the potential market relates; the competitors to the proposed data product; their strengths and weaknesses in relation to the proposed data product, and whether they are sufficiently funded to respond to any threat to the position of their product in the market.

Knowing the answers to these questions allows the business plan to be constructed and the proposed investment in the construction of the data product to be assessed. With the geoinformation data publishing market only in its infancy, many products collapse or have to be scaled down considerably after the business plan has been constructed owing to lack of market knowledge and the reality that the size of the target market is in many cases unknown. Many projects are run as pilot projects at this stage with perhaps sample data sets of one particular area being used in trials rather than the complete spatial coverage that is envisaged for the final product. The pilot project helps potential users to see the benefits of using the new product and assists them in then- understanding of its utility in their business environment. Once the market assessment is complete together with the business plan, decisions on which data are finally needed for the product, the design of the software front-end, the sales channels and the marketing exposure can be determined.

Intellectual property ownership: vital to success Electronic data products are made up of two types of intellectual property: the access software and the data. In data products there are usually many owners of intellectual property in one product. Both the software and the data can often be created as separate entities and the data can be contributed by many parties: it is only the act of publishing the data that brings the products together. Data publishing differs from normal text-based publishing; it is, however, similar to the book packaging industry. In the text-based publishing industry, the normal business relationship is between the publisher and the author of the work. It is assumed, unless the publisher is advised otherwise, that the author has created an original work. In cases where a few pieces of artwork have been used which have not been created by the author, the publisher requests the copyright holder’s permission to use those pieces of artwork and often pays a small fee for their use. In text-based publishing, the author or creator owns the intellectual property. Until recently it has been usual for the author to assign the copyright to the publishers for the term of copyright or for a fixed period, and hence the publisher has been licensed a substantial part of the value chain of the product. Authors are beginning to understand that not only can text-based publishing enhance their reputation but also that the work they have created can be of economic value for the future licensing of rights, either text-based or electronic media rights. As a result, many authors are requiring to retain the copyright in their work and hence retain the value in their work. Ownership of as much of the value chain as possible is the key to successful and profitable data publishing; hence this means owning as much of the intellectual property in the data product as possible. Trends are emerging where data publishers are creating

90

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

their own intellectual property for products and are having to use only minimal external resources to create the product and hence pay only minimal royalty payments to third parties. Even though the initial investment in this strategy is high, the publisher owns the majority of the value chain in the product. As long as the market price for the product does not drop during the development phase, the publisher will be successful. Also, if the publishing house owns the intellectual property rights, then it is able to license the data for use either in its own products or license others to use it, thereby receiving revenue for data that was created for another use. The relative value of a particular data set is often difficult to quantify but the publisher must determine which data sets will add most value to the product. In a spatial data panEuropean statistical product, for instance, it is necessary for the publisher to assess in which part of the product the greatest value lies. Only after this has been decided is it possible to negotiate successfully with data owners, as it is then known how valuable the data set will be to the end product. Is the value in the background data that give the spatial reference or is it in the statistical data? Naturally, in such a product as a spatial data panEuropean statistical data set the background data is essential to the statistical data but the quality is not of vital importance. It is not the quality of the background data that determines whether a customer purchases the product; it is the quality of the statistical data, the market focus of the chosen statistics and the quality of the presentational and analysis software front-end that determines the buying decision of the customer. Besides the publisher having to assess which types of data are of greatest importance to the final product, it is strongly advised that the publisher decides, prior to negotiation with data owners, which data sets are vital to the project. Using the example of the spatial data pan-European statistical product again, the success of the commercial product will depend on successful negotiation with the larger, more economically important nations of Europe. Even though smaller countries may wish to become part of the product, a spatial data panEuropean statistical product without significant data from France, the UK and Germany is simply not commercially viable. In the case of such a project, it may not be obvious at first with whom one should negotiate a country-wide deal; discovering who owns the intellectual property in the data is critical. It is the responsibility and liability of the publisher to ascertain that the organisation which states that it owns the intellectual property rights in the data actually does so. It is disadvantageous, i.e. expensive, to negotiate with a data owner once the data are embedded in a product as obviously the publisher has lost the opportunity to discuss whether the data set should be included. The spatial unit at which the data are available differs across Europe. For most countries there are central organisations such as national mapping agencies or national statistical offices with which the negotiation for obtaining the use of their intellectual property can take place. However, in Germany these organisations do not exist with the same remit and all negotiations must take place at Länder level.

Pricing of intellectual property within the product As discussed earlier, the aim of most information publishers is to own as much of the value chain in electronic products as possible. Most new publishing contracts now clarify the ownership of electronic rights for material that perhaps is currently destined to be a normal textbook. As technology advances, permissions for use of material in electronic form will become the norm, and by clearing the rights with the creator of the intellectual

EUROPEAN GEOINFORMATION PUBLISHING

91

property at the original signing of contract, the administration of the publishing house will be reduced and the publisher will have greater freedom to consider innovative use of licensed intellectual property, as suitable technologies advance. Interestingly, publishers of software tend to view intellectual property differently from information publishers and hence approach the licensing of their products from a different viewpoint. Software publishers often develop strong application programming interfaces (APIs) as part of their base products and encourage developers to build applications on top of the base software. The intellectual property created in the application resides with the developer together with the opportunity to exploit the product commercially; the underlying technology is licensed on a royalty or fee basis from the software house. This approach is taken by large software houses such as Microsoft in order to increase the use of their base technology. Commercial data products must be profitable. The greater the number of organisations sharing in the value chain the less the profitability of the product, unless the organisations are also investing in the product and reducing its risk. Unlike the model for negotiation of text-based contracts, which in the UK is regulated by the Society of Authors and in other parts of the world by equivalent bodies, there is no model yet established for the electronic data industry. Policies differ from country to country. For many organisations which are not used to levering extra income from a data set collected for another purpose, an internal policy change may be needed within the organisation before any negotiations for third party use of the data can occur. Key decisions must be made by the owners of data before entering into a negotiation:

¡ Is it important to them for their data set to be included in the product or are they happy to be excluded if the financial negotiation collapses?

¡ What is their policy for distribution of data? Do they view their data as being a revenue generator or do they wish more people to use their data set?

¡ How should they assess the value of their data? Taking the spatial data pan-European

¡

statistical product again, perhaps the country is small but they are data rich. Should they assess the value of their data to the product in terms of the area of their country or as a percentage of the total megabytes of the total product, or should all data providers be paid the same royalty independent of relative data volume, data quality or importance of the country? Finally, how might this product, which has no precursor and hence an unproved market, impact on any national product sold by the organisation?

These are all unresolved issues in many European countries. For data held by government agencies different historical imperatives and traditions of government structures often determine the decision regarding the licensing of data. Some have clear objectives of copyright protection and cost recovery and others are still trying to decide their policies. Until products have started to emerge and standards have been set, the number of data products, especially cross-boundary products, will remain small. Getting full agreement to exploit the intellectual property of government and private sector organisations is still, for many, a very difficult decision with no predetermined guidelines.

Importance of the software front-end The front-end functionality for European geoinformation data products is vitally important. This must be determined in terms of both how the product is expected to be used by the customer and the technical limitations of the customers’ hardware. If the majority of the

92

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

customers have 16 Mbyte of RAM in their machines it used to be commercial suicide to use a front-end that requires 32 Mbyte of RAM to operate the program efficiently. However, with the dramatic drop in the price of RAM this type of consideration is of less importance. Also if the product is priced at £1,000 then many organisations are happy to purchase a machine on which to run the software; this is not the case for the £50 package. Factors to be considered when designing the software front-end include the following:

¡ Should the product be able to be used within a fully functional GIS, and if so which fully functional GIS should be chosen?

¡ Should a generic front-end be included which would allow non-GIS users to browse ¡ ¡

the data and undertake simple analysis of the data and perhaps extend the publisher’s market into other non-GIS market areas? Should facilities be provided so that new data can be imported into the product? Should the software allow the data to be downloaded from the product and used in another application or should the data be encrypted so as to prevent this?

These are all questions which must be resolved before the front-end software is developed. For many products, the software front-end is developed by a third party and licensed for use within the product. Care should be taken to ensure that the software front-end can be used under the terms of the agreement irrespective of the volume of the product and also in new editions, if desired, under the original terms and conditions. It is very difficult to renegotiate a contract once the product is on the market and is more successful than expected or when a new edition is needed.

Language issues: vital to successful European products For a European product to be successful it is essential that language issues are addressed. Even though English is understood by most technically based Europeans, the success of the product may depend on the product being presented in more than one language; the preferred languages being English, German, French, Italian and Spanish. The product will be difficult to sell if it is not available in a language that is familiar to the end user. This is even more vital if the data product is to be used in universities: European students tend to read tertiary level textbooks in one of the five languages and so will accept the product as a learning tool if it is available in a language that they normally use for education purposes. Language issues are becoming easier to overcome. Many pieces of software have the tools for language embedded in the software and so localisation of the software is much easier than it was just a few years ago. For a CD-ROM-based European data product, it would be normal for the software to be in several languages on the disk. However, where the documentation is concerned, it is more economical for the publisher to print several translations of the text and ship only the appropriate language version with the product. Online help on the disk is becoming popular instead of printed documentation; this reduces the costs of production for the publisher and reduces the time to market as no printing time is required.

Production, brand and marketing issues Once the market for the product has been determined, decisions on the delivery mechanism of the product and hence the packaging of the product must be made. It may be decided

EUROPEAN GEOINFORMATION PUBLISHING

93

that the product is to be delivered by an on-demand service which results in the customer receiving a floppy disk or a CD-ROM. Alternatively, it may be a consumer packaged product for selling on a shelf of a retail store or it may be delivered by the Internet and hence a suitable home page will need to be designed. If it is to be a packaged product, the packaging can be simple and inexpensive if the product is needed by the market and the market is not used to seeing highly engineered packaging such as glossy covers and fancy boxes. Trends are certainly moving towards simple packaging even for £500 products. This is the reverse of the book publishing industry where poor paper quality and poor packaging is unacceptable to the consumer of an expensive, specialist, short-run book. Brand management of the product is important. A strong brand identity, either with related products or one created for the product which can be carried through on all marketing literature, can determine whether the product is a success or a failure. If data have been supplied by a well known organisation, it can often be beneficial to ask permission to use the organisation’s logo on the packaging or on the relevant delivery mechanism: this gives the consumer confidence in the product. However, care must be exercised by the publisher that the target market really does value the brand of the organisation; research must be carried out to assess that no sector of the potential market is at odds with the endorsing organisation. The customer base needs to have confidence in the product. This confidence is often transmitted to them via three messages:

¡ the product is being published by an organisation known for other quality products or services;

¡ the product contains data from valid sources which have undertaken extensive accuracy checks on the data;

¡ the product is not a one-off. Transmitting to the customer, before the first purchase is made, that there is a regular update policy that can be bought on initial purchase seems to give the customer reassurance. It is also satisfactory to the publisher as it can start to run the product as a subscription product and hence receive money regularly as well as reducing the financial risk.

Problems and pitfalls For pan-European data products there are some well known problems and pitfalls which must be identified and resolved at the planning stage. How they have been resolved must be discussed in the documentation of the product or in the online help systems so that the customer does not make assumptions about the data that are inaccurate. There is no formal CEN standard for pan-European data collection but there are several de facto standards which should be referred to during the data collection and amalgamation stage. These include:

¡ how the data can be matched across countries for comparison in terms of the scale of the data and their attributes;

¡ how the data sets can be matched at the boundaries of the countries; ¡ how the accuracy statements about the data can be verified. These decisions are vital, and how they are handled by the publishers will determine if the product is to be deemed credible by the end users of the data product.

94

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Sales channels for data products The sales channels for data products depend on the market for the final product. Market segmentation can be very complicated, but, taking a simplified model, it can be categorised as:

¡ ¡ ¡ ¡

the single or low volume markets; the high volume markets; the vertical markets; the new technology access mechanisms.

These markets are differentiated by several factors including:

¡ ¡ ¡ ¡

price; packaging; channel; delivery mechanism.

Products in the low volume market are often highly specialised and normally highly priced. The market is often the same market for which the original data were developed and so customers can draw parallels of use, such as an application in a similar industry. The holder of the original data or a small agent acting on the owner’s behalf is the publisher of the data, and the purpose for exploiting the data is normally to reduce the commercial risk to the holder of the original data in its further development. Care must be taken by the original owner not to give the purchaser similar strategic advantages to those they themselves have gained from the data: often the available products are aggregate data sets derived from the original data. The marketing and selling of low volume products tends to be simple. Clients are identified and often approached before the product is released—in many cases the money may already be in the publisher’s bank account before delivery of the product has been finalised, thereby completely eliminating the risk to the publisher. Sophisticated channels and agent’s fees are unnecessary for the selling of specialised products as customers can normally be identified from market knowledge. Packaging of the product can be very simple. If the product is highly priced but needed by the market, the market will be happy for the product to arrive either by file transfer or as a limited-run CD-ROM. Documentation is essential but again it need be no more than a good quality word-processed document or online help on the disk. For a product aimed at the high volume market, it is important that innovative sales channels are developed. It is crucial that the suitability of the product for the market is researched as the high volume market is expensive to access and very few products destined for this market are successful. Unlike traditional volume products such as books or music CD-ROMs, the volume retail outlets for the data industry are still in their infancy. Consequently, if adequate volume is to be achieved, less traditional methods of reaching customers and getting them to pay for the product must be introduced. The following are just a few ways of reaching the customer:

¡ Traditional direct mail from the publishers to names on well maintained, performing mailing lists.

¡ If the product is available in certain data formats, arrange for the vendors of the software products with data in that format to sell the product: then- channels are targeted and also tend to be data hungry. This may require some adjustments to be made to the

EUROPEAN GEOINFORMATION PUBLISHING

¡

¡ ¡

95

packaging of the product to reflect the involvement of the vendor and may require the other data formats, except for the proprietary format and the software vendor’s data format, to be hidden on the product. If the product is packaged as a CD-ROM this is technically very easily undertaken. Have the product cover-mounted on an industry magazine. When loaded onto a computer the product might reveal a demonstration presentation and then a telephone number or e-mail address where a software key can be obtained to unlock the remainder of the disk. The publisher would charge the customer when the software key is requested and will also be able to collect address details of the customer for future targeted mailings of other products. Advertise the product on the Web; the product could be downloaded to the customer once a fee has been paid to the publisher. In Europe, some traditional bookshops and computer stores are beginning to collaborate with publishers to distribute their new data products. However, as considerable shelf space has to be given to equipment and demonstrations, and considerable skills training has to be undertaken, the publishers are being asked to assist the bookstores by providing the hardware to run the software programs and to provide large discounts even though the shop is often not stocking the product but just giving the sales lead to the publisher. It is still unusual to find geoinformation products being given exposure in such retail outlets. The geoinformation data publishers are all trying to assist these outlets to work with the industry, but, as more consumer multimedia products come onto the market, it is increasingly unlikely that the outlets will give space to the geoinformation products.

A product targeting a vertical market sector is relatively easy to promote as long as it is an appropriate product for the target market. Even if the publisher has little previous expertise in a vertical market, it is possible to undertake successful business-to-business marketing and sales techniques. Identifying trade publications that people with buying power in the targeted business sector regularly read, or buying mailing lists associated with the chosen vertical market and mailing each name are both standard ways of reaching the vertical market customer. Care must, however, be taken to ensure that the images portrayed in the direct mail piece can be understood by the person receiving it and relate to their own work environment. Building a database for future mailings based on the response to an initial direct mail campaign using purchased lists is vital to the success of products in the vertical markets. Part of the added benefit of any direct mail campaign using purchased lists is the value of each new name that can be added to the database of the business. A chosen vertical market may have a trade association. It may be cost effective to form an alliance with the trade association to promote a product directed at the association’s market. The business model needs to be assessed by both parties; it can be as basic as using the mailing list of the association’s members in return for a fee, or it may involve forming a joint venture alliance where the trade association assists in the promotion, marketing and selling of the data product and in return receives a royalty on the sales made through its channel or even all channels. It must be decided whether a particular trade association reaches all the targeted market or whether it gives access to only one sector of the potential market for the product. New technology such as the Internet will become increasingly important as a sales channel. Currently it is little used for data publishing except as an advertising medium on the World Wide Web. As transaction security on the Internet improves, it is envisaged

96

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Figure 7.1 Proposed model for data publishers.

that the medium will become a major sales tool for data publishers. The Internet technology to view and analyse geographical vector or raster data is already available from the major software houses, but there has as yet been no major implementation of such technology by a major data publisher which changes its current sales channels. The technology normally consists of a server, authoring software and Internet viewers which are either downloaded free of charge from the Internet by the end user or, if used in an Intranet set-up, can be purchased for a minimal charge from the software vendor. Figure 7.1 demonstrates how a data provider could in the future gain added revenue from selling on the Internet and hence increase the value of the data industry. This model demonstrates that there are two ways for the data publisher to increase income using the Internet:

¡ by direct connection with a customer wishing to purchase online data or a packaged data product; and

¡ by use of the data in the Web page of a third party. Currently, the customer has either to travel to a retail outlet or to visit the data provider to view or purchase online data or a packaged data product. The data publisher probably has to pay approximately 35 per cent discount to a third party administering the retail outlet and may also have had the cost of installing and maintaining equipment in the outlet if an online on-demand service is operated rather than a packaged product service.

EUROPEAN GEOINFORMATION PUBLISHING

97

Using the Internet, the customer could access the Web page of the data publisher and download a Web browser to view the data. If the data set is large and perhaps sold as separate layers or as tiles of information, then these layers could be viewed by the customer. The customer could then make a buying decision and, by a simple e-mail request, buy the appropriate data which could either be delivered on disk or be provided over the Internet. This type of transaction would be considerably more profitable to the data publisher as the price to the data publisher would not have been discounted and, if the product was dispatched over the Web, there would be no packaging or other associated costs. If the customer decides, however, not to buy the product, it is possible for the software to record who accessed the Web site, what they viewed and the duration of their browsing the Web page. Direct marketing can then be used to introduce the customer to the further benefits of the product and also to market associated products to them. It has been mentioned that using the Internet can procure further income for the data publisher by use of the data in the Web page of a third party. The data set may be of use to a third party in order to add value to their Web site. A good example could be the Web site of a hotel chain or a car rental company which could in future put digital maps into a customer’s Web site to be used by the customer when choosing the most appropriate hotel/car to use during, say, a business trip. In the case of the hotel chain, the data could be authored into the Web site to show the location of the hotels on a country map. It would then be possible to search on a town location which would bring up a larger-scale map indicating local sites, street names and other places of interest. The customer could continue to pan-in; clicking on a particular hotel would bring a display of the hotel’s facilities to the screen. It would even be possible to make a ‘walking tour’ on screen of the facilities. Each time the Web site is entered and the maps accessed, a software log could make a record so that the data provider could be paid a royalty based on the number of times the data are accessed.

Pricing of data products Market pricing for data products is still very difficult to achieve as so few products are on the market for the geoinformation industry that the industry has not yet been able to determine the market price of products. Currently price is often determined by using a multiple of the following:

¡ the royalty being paid to the creator of the software front-end of the product; ¡ the royalty being paid to the owners of the intellectual property in the product; ¡ the production cost of the product including the packaging and the preparation of the documentation. The pricing can be staged depending on the target audience and the service that is being supplied by the publisher. It would be normal to have a commercial price, an educational price and a further supplement on top of these prices for update services, subscription services or perhaps user community newsletters. There are some interesting pricing examples of commercial packaged products which demonstrate that certain data publishers do not believe that Europe should be classed as one market and have one pricing structure. The same product is five times more expensive in certain countries and is being sold as a luxury data product; in others it is being supplied as an impulse purchase and is even bundled with new computers. Perception for the use and need of the product influences price and demonstrates that Europe is not yet treated by all data publishers as a homogeneous market.

98

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Skills required to develop pan-European data products The skills required to develop pan-European data products for the geoinformation industry are different from those normally found in a publishing house: they are in fact nearer to those needed in a book-packaging house. Unlike with normal product from a publishing house, the content provider and producer are the publishing house which has licensed the third party rights in the product. Its staff need the following skills:

¡ the ability to identify suitable intellectual property for the product and negotiate and contract successfully for use of the intellectual property in the data product;

¡ to understand user interfaces and how the customer will wish to use the product and then brief the software house appropriately;

¡ to understand the constraints of the data and produce the product within the relevant confidence limits of the data;

¡ to understand how to package the data successfully and deliver it to the customer in a form that the customer accepts as standard;

¡ to understand how to run non-standard sales channels. A lack of suitably trained staff and the lack of suitable training courses are affecting not only the European geoinformation data market but also the multimedia industry which is facing similar problems.

The future The future success of the volume European data products market is still in jeopardy owing to the immaturity of the market. The products are very expensive to construct and produce, and the market for many of them is still unproven. Many publishers are reducing their risk by inviting financial involvement from several partners in the product; however, the business plans often do not reflect the eventual investment of the organisation as it is still difficult to estimate the amount of work needed to produce a final product. As more organisations begin to understand the value of their data, it is becoming in some cases more difficult and in others easier, to negotiate a commercial arrangement for licensing data into a third party product. Until the market becomes significantly larger, the value of the data will continue to be difficult to estimate. Organisations do not know the value of their intellectual property either in terms of the value to the product of including certain data sets or in terms of the value of the final product to the customer. In the multimedia industry, market pricing of the final product has been seen to be levelling out and most consumers have an expectation as to the price of the final product. At present there are no similar expected price levels for geoinformation products: such data can be tens of thousands of dollars or under fifty dollars. No doubt, understanding of the pricing model will occur in the geoinformation data products industry as the industry reaches maturity. The sales channels for promoting European geoinformation data products are currently expensive. As retail outlets become more obsessed with being the channel for consumer multimedia products, it is likely that they will exclude volume geoinformation data products from their product range as they will be classed as being too specialised for their customer base. Other, less expensive channels will have to be found, probably through the use of new technology; a good example that we are already seeing is data catalogues run on the Internet. These are operated by third party sales operations or

EUROPEAN GEOINFORMATION PUBLISHING

99

software vendors who charge publishers for creation of a Web page for the product and perhaps for inclusion of a demonstration of the data product. When orders are received by the sales operation over the Internet, the data product is forwarded in the most appropriate form to the customer and a percentage of the revenue is passed to the publisher by the sales operation. This type of channel is cheap to operate and can be highly profitable for the sales operation. European geoinformation data publishing is an emerging business. Its viability is still financially shaky and analogies can be drawn with the multimedia industry where smaller organisations have found it difficult to maintain sufficient resources to complete commercial projects and are often refinanced or taken over during the long product development cycle. For the geoinformation industry to maintain its strong growth into the less traditional markets now that the technology is becoming more accessible, it is vital that the industry encourages and offers investment to the geoinformation data publishing industry. The European geoinformation data publishing industry is needed for the growth of the European GIS industry into the twenty-first century.

NOTE The views in this chapter are those of the author and do not necessarily reflect the views of any past or present employer.

CHAPTER EIGHT

Data integration for commercial information products: experiences from the EC’s IMPACT-2 programme ROGER LONGHORN

Introduction Geographic information occurs in many different guises and has widely varying uses, large numbers of data collectors and information providers and millions of ultimate end users, from the sophisticated subject specialist to the layperson looking for useful holiday information or better understanding of where his or her local tax money is being spent. It is not the purpose of this paper to examine the numerous fields of endeavour in which GI plays an important role. Rather, this chapter examines the problems that face producers of information products in their attempts to combine GI with other types of information to create a product for various commercial purposes, ranging from special consultancy service providers through to tourism products destined for bookshop shelves. The chapter looks at data integration problems from both market related and technical viewpoints. In the 1992–95 IMPACT-2 programme (Information Market Policy Actions—2) of the European Commission’s Directorate General for Telecommunications, Information Market and Exploitation of Research (DGXIII), GI was recognised as an important niche in the overall market for multimedia products. Market studies produced by various research organisations in the early 1990s predicted substantial growth in the market for GIS (geographic information system) hardware, software and related consultancy services (Dataquest, 1994, 1995; Frost and Sullivan, 1992, 1994; Cambashi Ltd, 1993). However, the IMPACT-2 programme planners were more interested in seeing how GIS, as a tool, could be brought effectively to bear on producing a range of multimedia products using GI and destined for the European and global information marketplace, rather than just traditional users of GI. The IMPACT-2 programme pioneered a two-phase approach to launching pilot and demonstration projects. In the definition phase, the 180+proposals received as a result of an open call for proposals published in the Official Journal of the EC were evaluated by a team comprising both GIS and media experts from outside the Commission but under Commission guidance. Based on the evaluators’ recommendations, the IMPACT Programme Committee (IPC), comprising delegates from all EU member states, selected 28 projects to receive partial funding up to a limit of ECU 50,000 to develop a prototype 101

102

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 8.1 IMPACT-2 projects—using GIS to create media products

product and business plan to be submitted for a further round of evaluations at the end of six months. From this second round of evaluations which included prototype demonstrations again conducted by external experts under Commission guidance, eight projects were selected to receive partial funding for an implementation phase, each receiving up to a limit of approximately ECU 300,000 (actual sums varied depending upon the mix of partners in the project consortia). A report of the projects selected and their widely varying media product themes is available from the European Commission (IMPACT Central Office, 1994). This chapter describes the data integration problems which were met and (usually) overcome during execution of the full implementation phase of these final eight projects, listed in Table 8.1. The themes of the final products span a wide range of subject areas and potential markets, from secondary schools to technical consultancies, from business and local government users to tourists. Interestingly, the products destined for more sophisticated end users presented fewer problems in data integration than those for a wider, mass audience. Full-colour glossy fact sheets (40 pages) regarding the IMPACT-2 programme and the product goals of each of the eight projects are available from the European Commission (IMPACT Central Office, 1996a). Specific details on the individual projects are brought out in this chapter only when relevant to the data integration issues being discussed. Typically, all projects prepared detailed working papers identifying the target market(s), data provision, data processing (software and presentation techniques), overall system development, visualisation and delivery methodologies, pilot data plan and data integration plans.

Availability and integration of Gl into information products In the instructions to proposers to the IMPACT-2 programme, special emphasis was placed on the problems that proposers should expect in acquiring the data needed to produce their final products. This was based on experience gained by the Commission’s IMPACT-2 programme management team during an earlier call for proposals (CfP) for similar projects using interactive multimedia tools, launched just prior to the Information Services Based on GIS Technology CfP. These 22 multimedia projects are fully described in a separate set of fact sheets available from the Commission (IMPACT Central Office, 1995). Availability of data may be considered by some as being a separate problem to that of integration of data. However, this author sees the two as inseparable, especially in the

DATA INTEGRATION FOR COMMERCIAL INFORMATION PRODUCTS

103

common case in which multiple data sets are given added value by a product developer, which is the entire rationale for producing the product in the first place. If even one important class of data is missing, the entire product may fail commercially as being unacceptable to the target market. It should be noted that the problems reported in this chapter beset projects which had already completed a six-month funded definition phase. During this period, detailed market analyses were undertaken by the project partners and working models or similar product prototypes were completed and presented to the expert evaluators in order to secure the second round of funding. In other words, the project partners, including professional, knowledgeable and commercial firms, were embarking upon an information product development trail which was well researched and for which potential problems and pitfalls were supposedly well understood. Of the eight projects listed above, only three fully completed their development schedules, including all the data planned for their product launches, within the two-and-a-half year project period of the original IMPACT-2 proposal. It is important to keep in mind that the final eight IMPACT-2 sponsored projects covered quite different end-user markets, both high and low volume and high and low cost. They included:

¡ one inexpensive environmental education product aimed at secondary schools; ¡ two professional products destined for end users who would be experts in their fields, of high value and high cost to the end user;

¡ three tourist/travel/city information dissemination products for general use by laypersons, ¡ ¡

either on a low cost basis to travel agencies or typically on a ‘no pay’ basis to the ultimate end user (the tourist or traveller); one real estate agent/local taxation oriented product, to be used in office environments, of moderate cost (less than ECU 2,000) and requiring data from local government and business; and one general information GI metadata product incorporating ‘information about geographic information’, of low cost or perhaps even no cost to the end user.

Therefore, the problems faced by these projects and the methods used to overcome those problems are pertinent to hundreds of similar new product plans which are even now being developed in the information marketplace (see Chapter 7).

Main problem areas for GI integration The projects discussed here faced many data integration problems, not only because some required the integration of complex data from across national boundaries, but also because multiple data sources were required in nearly all cases even within local areas, such as a city. The data to be integrated into the final product also ranged from fully georeferenced information (street networks, etc.) to totally non-geographically referenced text and images (tourism information). Developers were usually faced with the problems of acquiring data sets which they did not always control or own. Most of these data needed to be geo-coded for presentation or analysis purposes crucial to the success of the product. All of these general problems manifest themselves in different ways at the data integration stage of the project. Specific problems are described in the following sections. The problems fall into two basic classes: market-related (including legal/intellectual property rights (IPR)) issues and technical issues.

104

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Market-related data integration problems Market-related problems include such considerations as negotiating legal rights to use data from multiple sources, data holders’ perceived value of their data prior to and after integration by the producer, timing of data acquisition and data integration in relation to product release schedules, working within multiple data access policies across multiple data providers, completeness of data, and working with ‘first time’ data collectors.

Producing a product for which the data cannot be legally sold This intriguing problem beset more than one project, especially from the wider list of 28 projects originally selected for definition phase work. In several cases, the most readily available data set which would satisfy the needs of the project was freely available, but with the stipulation that the data themselves could not be resold for commercial gain. This was especially true of several interesting product proposals focusing on environmentally related projects. For example, at the time of the IMPACT-2 programme, the CORINE land cover data sets for the whole of Europe, collected at considerable cost, spread equally between the European Commission and the member states, could be acquired at ‘cost of distribution’ but could not be resold commercially, even after adding value by a third party. In the case of the ENVIDUCATION project, the final product needed to be very inexpensive to sell into the secondary school environment. The project was eventually successful in acquiring the GRID (Global Resource Information Database) environmental data for Africa from the United Nations Environmental Programme (UNEP), but under the UNEP terms and conditions which prevent the data being resold (see also Chapter 3). Therefore, ENVIDUCATION took the quite practical approach of developing the data handling software to be used in the classroom, for doing environmental education project work, and putting this on one diskette for distribution and sales. This diskette was then accompanied by the GRID data, on a second diskette, available ‘at no extra charge’ to be used by the software. In theory, other data sets could be used in the future by the classroombased software.

Unplanned data integration costs lead to product price increases Another problem related indirectly to data integration is that of higher than expected integration costs leading to higher than predicted retail prices for the final product. In the worst case, this can spell the end of the product. In the best case, this almost certainly reduces the profit margin available to the product producer. The added cost of data integration in multiple sourced products affects not only original acquisition cost but also future updating and maintenance costs.

Timing of data collection and product updating Timing of the data gathering and updating tasks across multiple partners and multiple data sources in order to hit the market at appropriate times with both initial product launches and with future updates is a difficult problem to overcome, and is directly

DATA INTEGRATION FOR COMMERCIAL INFORMATION PRODUCTS

105

dependent upon the data integration plan to be implemented. In order to price the final product accurately, the designers must know as much as possible about all aspects of the data sets destined for the final product. This includes the acquisition and update frequency for all data sources, estimates of the amount of data that will change over time, estimates of how much additional effort will be needed to incorporate change into existing product data, and risk analysis of what will happen to the end product if the updates do not appear on schedule.

Value of the data set as seen by the holder underestimates cost of integration for the producer Some data collectors/holders are only now waking up to the fact that they sit on a valuable asset. For example, a bus company or railway operator spends considerable sums on setting timetables and promulgating these to the public. They are now realising that this information is not only important to their daily operations and planning, but is also an asset of value in its own right, on which further income might be generated either by themselves or via third party value-added activities. Unfortunately, the cost for such data quoted to the third parties can vary quite substantially depending upon how well the data owners understand the information market. However, lack of awareness by single-source data holders of the many steps involved in integrating their data into a final product can make acquisition negotiations by the producer more complicated or even result in removing the data set from the planned product. This adversely affects both the producer and the data holder.

Multiple data access policies across multiple data providers The MAG IS project pilot depended upon being able to negotiate access to multiple data sets held by the 250+ communes in Denmark. A major partner was the company which provides computer services to those communes. Yet it transpired that there was no common data access policy exercised by the individual communes. Some offered data on an almost free basis (but without guarantees) while others asked quite substantial royalties (for various levels of guarantee of quality, completeness, etc.). Even data collection and storage formats were found to differ markedly across multiple communes for similar data sets. Historically, there had been no requirement to standardise, and only newly collected data were now being collected to an agreed standard.

Completeness of data sets Databases of urban data, often available from local or national government, are equally often not up to date and/or of uncertain quality, comprehensiveness and completeness. This is a data integrator’s nightmare. The product developer may be forced to radically alter the product’s content plan after the development project has begun, based on more detailed discussions with actual data providers. It is important that the developer ask the difficult, penetrating and concise questions regarding all aspects of the data on offer before finalising development plans, time scales and budgets.

106

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Overcoming data collection problems caused by ‘first time’ data collectors Data which are potentially very important and useful to information product developers are often collected and/or digitised by ‘first time’ users, for their own requirements. Since these users are oblivious to the developer’s existence and know nothing about its product plans, they have little interest in collecting exactly the same level of data that the developer might wish and/or in a standard format. Thus, it falls to the developer to incur the cost of format conversion, of adding additional data items, of quality checks, etc. Liaison with potential data suppliers very early on in the planning cycle is well advised, especially if the developer’s expertise could help the potential data provider to do a better job for their own proposed uses, even without considering possible further use/re-use of the data by a third party. This could be a role for major trade associations or multinational companies, rather than the single product developer. How does the data owner see the developer—potential revenue source or competitor? Another problem related more to data provision is lack of cooperation by data holders who may see the developer’s product or service as a competitor—even where their own mandate (public or commercial) to provide services precludes them from directly competing with the developer viz à viz information products. The data holder probably does not have access to the other data sources that the developer is integrating into its new product or service. In most cases they will have neither the mandate nor the additional resources needed to collect or acquire such additional data, let alone to integrate it into a more complete information product. Nevertheless, certain data holders fear releasing their data to third parties who are seen as potential competitors. Unfortunately for the product developer, this stance often becomes known only after initial discussions have occurred during project planning stages which looked ‘very promising’. Let the developer beware—until data supply contracts are actually signed and executed, those data do not exist for the purposes of the project. Overcoming lack of data at least cost As one might expect in pilot projects, a major problem in completing the products was often lack of certain necessary data. This problem was often known at the time of the project proposal and provision was made to collect the required data (for example, a full city street network in the case of one project). Plans to acquire the required data from potential existing suppliers failed to mature because of the high cost being asked and incompleteness of the data set available even for that cost. A decision was then made, after project budgeting and start-up, to collect the missing data within the project framework itself, perhaps using partners located in the relevant region, city or country. Unfortunately, the workload involved in doing this often exceeded the planned budget, and the region, city or country in question would eventually be dropped from the pilot scheme until such time as all the relevant information could be gathered. Technical data integration issues As might be expected, most of the technical data integration issues focused on standards (of many types) and procedures for attaching geographic references to attribute data.

DATA INTEGRATION FOR COMMERCIAL INFORMATION PRODUCTS

107

There were also specific problems encountered working with cross-border data and in meeting users’ presentation requirements, especially across a wide range of user levels of sophistication.

Data formats are far from standardised Many projects, especially those being carried out by less experienced developers, found that the wide variety of data formats in use in the GIS market was a major obstacle (IMPACT Central Office, 1996b). Not only do nearly all GIS software platforms use unique and proprietary internal data formats, but also the data collected by thousands of different users, for thousands of different purposes, are often collected and/or stored in yet other in-house formats, for reasons that may be peculiar to each individual project or intended use. It is interesting to note that several of the IMPACT-2 projects developed yet further internal formats in order to deliver their final products. This was often done in order to encapsulate data of different formats from multiple sources into a single, common internal format which was acted upon by the GIS presentation/analysis tools. Especially where new interactive software was created specifically for the final product, rather than pay royalties or licence fees for standard, off-the-shelf GIS packages, the newly created format often followed no current standard, thus potentially creating yet more standards-related problems for the future.

Standardisation of codes used in data from multiple sources Problems in integrating GI can often extend into the presentation aspects of the data, as well as in specific definition of certain attributes, i.e. questions such as, what does a threestar rating mean in this guidebook compared with other guide books?, must be resolved. Even in something as well regulated as the travel industry, with a long service history, three stars in a French camping guide may not mean the same as three crowns in a UK or Dutch camping guide.

Working with cross-border data Georeferencing and displaying attribute data across national boundaries where the underlying map base is still not seamless is a problem for pan-European data. The EU RI P I DE S project produced a product destined for professional planning organisations, especially those with cross-border needs. This product was developed in cooperation with E U ROSTAT (the Statistical Office of the European Communities) and the national statistical bodies of various EU member states, to provide 46 socio-economic variables to NUTS-5 level (Nomenclature of Territorial Units for Statistics—to single census unit level) georeferenced to the SABE (Standard Administrative Boundaries of Europe) digital database produced by the MEGRTN Group (the marketing organisation of 18 European mapping agencies). SAB E boundaries still have some anomalies: for example, borders do not always match (see Chapter 2).

108

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Combining many different types of information af fects presentation technology One of the products arising from the IMPACT-2 CfP was OMEGA which collected and presented volumes of GI metadata (information about information) about maps on a CDROM. These data were originally to be presented to the users via a simple GIS interface, so that metadata could be searched and/or presented ‘geographically’. The complexity introduced in trying to use a GIS tool to accomplish this proved too cumbersome for the typical expected end user. Following initial user trials the project resorted to using a well established multimedia development toolkit to provide the user front-end, and restricted the level of complexity of data held on the CD-ROM-based product.

Attaching geographic references to attribute data In the TiTAN project, attaching textual travel information to an underlying mapping coordinate system took three times longer than originally planned. A motorist organisation’s requirements dictate how it collects data and uses it in the field and in printed guides. The accurate georeferencing of this data onto a specific map base is not needed in normal operations, either in the field (roadside assistance service) or in its many publications.

Accommodating multiple data types and missing data The VITAL project planned to develop complete city guides for the tourist or business traveller, focusing on four cities (Amsterdam, Brussels, Munich and Lisbon). These guides would be available via both kiosks placed at strategic locations around a city and via a software/data package for desktop computers which could be located in travel agencies, in the travel departments of large companies, etc. The project would build on detailed street network information already available from a major partner or still to be collected in the target cities where it did not exist. To these primarily line-related features would be added area-related data (for example, parks, built-up areas, water areas and industrial zones), as well as point information (for example, petrol stations, restaurants, hotels/ motels, railway stations and car parks). This location-specific information was held in textual form by another project partner, a major European motorist organisation. Because the various data types (street networks versus motorist and tourist information) were held in different formats by the different partners, the first data integration problem was to agree a common exchange format. This was developed within the project. A further complication arose because new data were also to be captured with regard to some of the location information. Thus, provision had to be made within the data capture and updating system of the motoring and tourist organisation in this regard and extended to the project’s proprietary exchange format. Location-referenced data such as house numbers attached to street networks were also missing from some data sets but present in others. This is partly owing to the data collection and maintenance (update) schedules of the major companies working in this area in Europe. Many such companies work only on a local basis (one or two cities) or a national basis (major towns and cities in one country), but very few work across Europe. Of the latter, the data collection effort needed to maintain accurate databases is onerous and expensive, hence the data are collected on an ‘as needed’ basis, to suit commercial and public clients. Therefore, the data needed for a particular city or area may not be available as and when needed.

DATA INTEGRATION FOR COMMERCIAL INFORMATION PRODUCTS

109

Handling specific types of data from the integration viewpoint: timetables It is one thing to make an image of a timetable page available on the screen of a tourist/ traveller information system—this involves either acquiring the original information in digital form from the transport operator or securing permission to scan the relevant page(s) and display as an image—it is quite another to try to attach the detailed schedule information to individual bus routes, by georeferencing subsections of the timetable to individually located bus stops, for example. While the latter approach was felt to be much more user friendly to the traveller who only sees the timetable information relevant to his/her location in the city, the added cost to the developer was probably not worth the effort. The developer should beware the knock-on effect of such features, especially when individual sections of a timetable may change independently of the whole schedule (for example, summer bus routes or train schedules which change according to school closings and tourism requirements). If the data integration plan does not consider these issues from the outset, the developer could face serious updating problems in the future, both in inserting the newly changed data and in making it available to end users. Some local transport information, such as bus timetables, were simply not available in some locations. Other timetable data were only available under copyright/IPR restrictions with high royalties demanded for both the initial load and for subsequent updates. Matching the update frequency of the scheduling information to the planned updating frequency of the overall product, per city, also caused concern, and had perhaps not been fully considered at the project planning stages. In addition, the providers of external data could not always be tied down to a long-term agreement on data collection, formats, who would pay for converting to internal project transfer formats, etc. The problem is that, for many data providers, such provision is secondary to their formal tasks and/or public man-dates to provide services. Thus they cannot be expected to alter their current work practices to accommodate a third-party requirement unless they are adequately compensated for the extra work—and, in some cases, even this is not sufficient inducement.

How to approach the updating and temporal aspects of GI products Once the initial products had been developed, or during development, it became apparent that some would need expensive updating activities if they were to remain of use to the end user. This was especially true of the travel/tourism guides, but also of other systems such as real estate applications or marine information. The specific data integration problem arises when one part or component of a complete data set needs to be updated more frequently than the others. In this case, the design of the whole production system must be geared to handle the situation. One method adopted by more than one project was to create a more neutral, intermediate data holding stage, in which data from multiple sources and of multiple types (map coordinates versus text, images, etc.) are held.

Technical problems in integrating multiple source GI Data were made available from multiple sources in multiple formats, often including yet further permutations owing to intermediate transfer formats used to exchange the data. Further conversions might then be needed before the multiple data sets could be integrated for use by the GIS which used their own proprietary data storage and handling formats.

110

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

All these intermediate handling stages added significantly to the overall cost of putting the final products together, and will also increase data maintenance and updating costs throughout their life.

Lessons learned—or not? Various lessons regarding the integration of multiple source GI from more than one European country were learned during execution of the eight IMPACT-2 projects, as can be surmised from the discussion. Future product developers are well advised to examine these carefully in relation to their own plans for new GI/GIS products and take note of at least two major lessons in planning their own GI related products.

1. Determine legal ownership of the data and agree legal use/re-use terms If at all possible, own your data sources. If that is not possible, then sort out ownership, sale and resale rights very early on in the project (see also Chapter 7). Projects in which one or more partners owned the IPR on the material planned for dissemination via the product caused few problems. As soon as more than two or three data providers entered the picture, IPR issues became more complicated. This was especially true if the providers were from different organisational environments, e.g. local government holding potentially private data versus a company with commercially orientated data. Problems often arose over determining a ‘fair price’ for the raw data, ranging from those data providers (often governmental) who were willing almost to give it away, to those demanding high royalties because they had a higher level of sophistication in understanding the value of their data, its collection and maintenance costs, and value to the end user.

2. Ensure that your production plan covers all data integration requirements If many independent data sources are to be incorporated into the final product, then pay great attention to the overall production schedule and technical plan, including added costs of all the intermediate processing and storage steps. These added costs can come as quite a surprise even to established publishers, unless they are already very experienced in the multimedia business.

NOTE The views expressed in this chapter are those of the author and do not necessarily reflect those of any past or present employer. REFERENCES CAMBASHI LTD (1993). Geographic Information Systems Markets and Opportunities: European Supplement. Cambridge: Cambashi Ltd. DATAQUEST (1994). Profiting in GIS Services: ‘Coopetition’ in a Converging Data and Services Market, Focus Report. San José: Dataquest.

DATA INTEGRATION FOR COMMERCIAL INFORMATION PRODUCTS

111

DATAQUEST (1995). Geographical Information Systems User Survey 1995. High Wycombe: Dataquest Europe Ltd. FROST and SULLIVAN (1992). The European Market for Geographic Information Systems (GIS), Summer 1992. Silicon Valley, Calif.: Frost and Sullivan. FROST and SULLIVAN (1994). World Geographical Information Systems Software and Services Markets. Silicon Valley, Calif.: Frost and Sullivan. IMPACT CENTRAL OFFICE (1994). GIS Project Fact Sheets—Final Definition Phase Report 27.4.1994. Luxembourg: IMPACT Central Office, DGXIII/E. IMPACT CENTRAL OFFICE (1995). IMPACT 2 Interactive Multimedia Projects, Fact Sheet. Luxembourg: IMPACT Central Office, DGXIII/E. IMPACT CENTRAL OFFICE (1996a). IMPACT 2 GIS Projects—the GIS Fact Sheets. Luxembourg: IMPACT Central Office, DGXIII/E. IMPACT CENTRAL OFFICE (1996b). IMPACT 2 GIS-3160 ENVIDUCATION Final Report. Luxembourg: IMPACT Central Office, DGXIII/E.

PART FOUR

Legal and Institutional Issues Associated with the Development of Multinational Databases

CHAPTER NINE

Legal protection of geographic information in the EU MIREILLE M.M.VAN EECHOUD

Introduction In 1995, the European Umbrella Organisation for Geographic Information (EUROGI) commissioned RAVI, the Netherlands Council for Geographic Information, to conduct a survey on the legal protection of geographical information in the EU. Initially, the situation in five countries was reviewed in a pilot study. Subsequently the remaining ten EU countries were researched. This chapter is based largely on the final report of February 1996 (Eechoud, 1996).

Why a study on legal protection? The mission of EUROGI is ‘to promote, stimulate, encourage and support the development and use of geographical information and technology at the European level, and to represent the common interest of the geographical information community in Europe.’ The development of an efficient and effective geographical information sector demands an adequate legal framework. This framework should facilitate the distribution and use of geographic information. Legal issues that merit attention are, among others, copyright/ intellectual property rights (I PR), liability, privacy, data protection and access/ commercialisation issues. The first legal issue that EUROGI has decided to address is legal protection of geographic information. As the European market for geographical information develops, so does the need for adequate and harmonised legal protection of geographic data and information products.

Aspects studied It should be noted that the study conducted focused on the protection of information and its components, data, rather than on protection of hardware and software used to collect, update, manipulate and retrieve geographic information. Furthermore, the words ‘information’ and ‘data’ are not used as distinct from each other unless specifically stated. 115

116

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

On asking what rights there are in information, one should start by examining copyright (droit d’auteur) and neighbouring rights.1 The term ‘neighbouring rights’ refers to rights that are akin to copyright. In addition, the law on unfair competition can in some circumstances protect producers of information against the copying of their work. Another type of legal protection is legislation that specifically protects geographic information or its producers. Even where (exclusive) rights exist in information, the owner of these rights does not have unwarranted powers with regard to the information product. Therefore, a number of possible limitations to (exclusive) rights have been studied. These include limitations embedded in the laws on copyright and neighbouring rights,2 the effect of public access laws on exclusive rights in information, and rules of competition that may restrict the use of exclusive rights.

Copyright and neighbouring rights Copyright is generally divided into two different types of right. One sort is labelled moral rights, which denote rights of the actual creator, described as the author, to oppose mutilation of his/her work and so on. Moral rights primarily serve the immaterial interests of the author. This chapter does not discuss these rights further. The other category of rights are economic or exploitation rights, also referred to as the exclusive right to reproduce and publish (or disseminate) the work. These are the rights that make it possible to exploit the product, and they are of utmost relevance in the geographic information sector. In the EUROGI study (Eechoud, 1996) attention therefore centred on economic rights. The copyright systems of the member states of the EU share common notions to a considerable extent. Basically, this is owing to the fact that there are longstanding multilateral treaties on copyright such as the Berne Convention of 1886 and the Universal Copyright Convention of 1952. More important for harmonisation today is EU legislation in the field of intellectual property law. Topics of directives issued in recent years include:

¡ protection of computer programs (CEC, 1991, not discussed here); ¡ harmonisation of rental and lending rights and certain neighbouring rights (CEC, 1992);

¡ harmonisation of the duration of protection of copyright (CEC, 1993); and ¡ legal protection of databases (CEC, 1996). International treaties on copyright As has been indicated, the two treaties that deal with copyright are the Universal Copyright Convention (1952) and the Berne Convention (1886, latest revision: Paris Act 1971). The Berne Convention (BC) for the protection of literary and artistic works is the most important copyright treaty; over ninety states have adhered to this convention, including all EU countries. The BC works on the basis of reciprocity: it obliges parties to give subjects of other BC countries the same protection as their own nationals enjoy under national copyright law, with at least the minimum protection that has been set out in this convention. Article 2 sub 1 of the Berne Convention states that subjects of copyright are works of literature and art, comprising every production in the literary, artistic and scientific domain, whatever may be the mode or form of its expression, such as books, pamphlets or other writings; lectures, addresses, sermons and other works of the same nature; dramatic or

LEGAL PROTECTION OF GEOGRAPHIC INFORMATION

117

dramatical-musical works; choreographic works and pantomimes, musical compositions with or without words, cinematographic works to which are assimilated works expressed by a process analogous to cinematography; works of drawing, painting, architecture, sculpture, engraving, lithography; photographic works to which are assimilated works expressed by a process analogous to photography; works of applied art; illustrations, maps, plans, sketches and three-dimensional works relative to geography, topography, architecture or science. As can be seen, geographic information is only one of many types of work eligible for copyright protection. Although extensive, the list is not exclusive, i.e. it gives only examples of works that fall within the scope of copyright. Similar listings have been incorporated in various national copyright laws. An essential demand to be met in copyright is that a work is original, which usually means that the work should show a creative/intellectual effort by the author. This requirement is not expressly stated in the Berne Convention, nor is it in many national copyright laws. It is, however, generally accepted that the concept of a work (oeuvre in the French text) in the sense of copyright implies originality. The interpretation of what ‘original’ is exactly, may differ from country to country (see later).

European Union regulation When the Council of Ministers of the European Union issues a directive, member states are obliged to implement it in their national legislation. This may involve the drafting of a new law or the adaptation of existing laws. Another possibility is that the national legislator shows that the existing law already complies with the new directive. In these cases changes in national law need not be made.

Directive on the legal protection of databases Recently, the Council and the Parliament of the European Union adopted the directive on the legal protection of databases, which all member states must have implemented by 1 January 1998 (CEC, 1996). Since the first draft directive was published in 1992, the plans for the protection of databases have changed considerably. Most notably, the directive no longer covers electronic databases only, and the proposed compulsory licensing schemes have been dropped. A database is defined as: a collection of works, data or other independent materials, arranged in a systematic or methodical way and capable of being individually accessed by electronic or other means, (section 1) Many geographic information products will be covered by this definition, since it applies not only to ‘modern’ electronic/digital files but also to collections of data in paper format (maps and plans, for instance) and on other media such as microfiches. As soon as raw data have been manipulated, the result appears to fall within the scope of the directive. Computer programs used for production and operation of the electronic database do not fall within the scope of the directive. However, parts of the database that are necessary for the operation or consultation, such as the thesaurus or an index system, may be covered by the directive. It would seem that in practice it will not always be easy to determine which ‘part’ of the database is a computer program and therefore not protected under the directive.

118

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

The directive obliges EU member states to grant copyright protection to databases on condition that they are original. Original is described as the personal intellectual creation of the author (section 2 sub 1). No other criteria may be set. More importantly, the directive also provides for a sui generis right intended to protect the producers of non-original databases. This so-called extraction right gives the producer of a database the right to resist retrieval and use of all or a substantial part of the data of the database. Not every database qualifies for this protection: the production, maintenance or presentation of the contents of the database must have cost a substantial investment. What substantial means in this respect is not specified. The right to prevent unfair extraction lasts for fifteen years on completion of the database. If the database is made public before the fifteen years from the date of completion have passed, the term of protection is fifteen years from the date of publication. Where a substantial investment is made (in time, money, expertise, etc.) that results in a substantial change of the contents of the database, the database is regarded as a new one, and another term of protection will start (section 12). It is thus possible that databases might enjoy eternal protection. Since it is common for geographic databases to be updated at regular intervals, producers of such databases might just gain long-lasting protection. In countries where databases already in existence before January 1998 are protected by copyright, even though they are not under the terms of the directive, the term of protection for these existing databases remains equal to the term of copyright protection (usually 70 years).

Directive on harmonisation of duration of copyright On 1 July 1995, the directive on harmonisation of the duration of copyright (CEC, 1993) must be implemented in member states. Consequently, as a general rule, in all European countries copyright will last until 70 years after the author’s death, or 70 years after publication in the case of anonymous works or works of which a legal entity (corporation, state, etc.) is the author. For most countries this will mean that the term of copyright will have been prolonged by 20 years, since 50 years was the most common duration.

National copyright laws Types of product protected As has been said above, national copyright laws in the EU are in general alike with regard to the types of work that qualify for copyright protection, typical products of geographic information such as maps often being mentioned explicitly as subjects of copyright. The essential requirement of originality is interpreted similarly in European countries, Ireland and the UK being notable exceptions. In Ireland and the UK it suffices if skill and labour are involved in the making of a work, whereas the continental concept of originality requires that the creative, intellectual efforts of the author(s) is perceptible in the work. This means that copyright protection stricto sensu is more readily available for geographic information products in Ireland and the UK than in the other countries (see Chapter 11). Many geographic information products are either (paper) maps and plans, or databases. In various countries, courts have decided that maps and plans can be original and therefore protected by copyright. It is recognised that the design of a map (that is to say colouring, typesetting, symbols used, etc.) can make it original. In the Netherlands the process of

LEGAL PROTECTION OF GEOGRAPHIC INFORMATION

119

generalisation has been recognised as one of the elements that expresses the creative/ intellectual effort of the author. Although collections of facts are, at least in theory, possible objects of copyright in any of the countries researched, the required originality may prove to be prohibitive (except in Ireland and the UK). In databases the original character must be perceptible in the selection and arrangement of the content. However, the functional and technical demands that the database needs to meet (for reasons of quality, accuracy, completeness, user friendliness, etc.) adversely affect originality. Thus the use of, for instance, technical standards (whether local, national, or international such as CEN standards) and standard classifications makes the product less original. Scope of protection Copyright encompasses the exclusive right to reproduce a work and to publish or disseminate it. The reproduction right envisages both the production of exact copies and of adaptations. The right to publish or disseminate includes the right to sell copies of a work or to exploit it through rental or lending. Following the EU directive on rental and lending rights (CEC, 1992), permission of the author is needed in all EU countries with regard to (commercial) rental. The situation with regard to lending is diverse, specifically concerning library privileges. It is true that, for all countries, once physical copies of an information product have been brought onto the market by or with permission of the owner of copyright, the latter has no more influence over the subsequent sale of these copies within the EU. Ownership of rights On comparing the situation with regard to ownership of copyright, one will find that there are quite substantial differences in national laws. The one basic principle that all copyright acts share is that the author, that is to say the actual creator of the work, is the initial owner of copyright. So it is not the person or company who finances and/or commissions the production of a product who automatically owns the exclusive right to exploit the work. Copyrights may, however, be transferred or given in licence.3 If several persons have contributed a creative effort, they will, as a general rule, jointly own the copyright. A question that seems to be of particular relevance to the geographic information sector is whether an employer automatically owns the copyright on works produced by his employees. In the Netherlands, the answer is affirmative, unless the parties involved have agreed otherwise. Greece has a similar system. In Germany, companies and other legal entities can never be the initial copyright owners. However, it is generally assumed that by accepting a job the employee implicitly grants the employer the licence to exploit the copyright on the works that the employee creates in the course of his duties. Belgium has yet another rule and so does Italy. Even without giving an exhaustive account of all the rules in different countries, it is safe to conclude that it is complicated to decipher who actually owns copyrights or has a licence with regard to a particular information product.

Special regulations in copyright and other acts Having discussed the protection of original works by copyright, we now turn to specific rights in information products that are not eligible for copyright protection. These special regulations may be found in copyright acts or in special acts.

120

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

The Nordic copyright laws protect producers through the catalogue rule. Catalogues, tables and similar compilations that contain a large amount of items may not be reproduced without the author’s consent. Other countries do not have similar regulation, except the Netherlands (but there, graphic products such as maps do not enjoy special protection). Besides specific regulation for non-original collections of information or data, nonoriginal photographs are also protected in a number of countries. This is true for Austria, Denmark, Finland, Italy, Spain and Sweden. As is the case with ownership of copyright, these countries have different provisions regarding who owns the right to resist unauthorised reproduction. In Austria and Italy, the employer (be it a physical person, company or other legal entity) owns the right on photographs produced by employees, whereas in, for instance, Denmark only natural persons have these rights. The copyright laws of Greece and Ireland provide protection to the publisher of a work, to the extent that his consent is required for the exact reproduction of the design (typesetting) of the product. Under Spanish law the publisher of previously unpublished works that are in the public domain (that is to say, of which the term of copyright protection has elapsed) has exclusive rights over the publication. In Sweden, Germany and Portugal, national/federal mapping agencies (NMAs) are granted exclusive rights on some of their products. In these countries these rights are embedded in laws that also regulate the activities of the NMAs. In particular, the Portuguese law seems to have been drafted without taking into account the existing intellectual property law (including copyright law). This might cause difficulties if rights stemming from this specific law collide with, for instance, copyright. In general, one might argue that the fact that these special rights are only available for certain producers and products, and that they can coincide with copyright and related rights, does not contribute to the transparency of the legal system.

Protection based on the law of unfair competition In all member states of the EU, freedom of competition is a basic principle. In broad terms, freedom of competition means that it is allowable to profit from someone else’s achievements. Thus, in economic life it is generally permitted to copy products, since this is deemed to serve technological and economic development. Copyright and related rights constitute important exceptions to this principle, since they grant exclusive rights on information products. If copyright or similar rights are not available, producers might turn to the law on unfair competition to protect their interests. The laws on unfair competition differ widely.4 In some countries (Ireland, UK, Netherlands) the unwritten law on unfair competition in practice offers no or very little protection against the copying of products, while in, for instance, Germany, Belgium and Austria, the copying of information products is in some circumstances actionable under the unfair competition acts (Court of Appeal, Frankfurt, 1995). That protection against copying based on unfair competition law is not easily granted may become clear from the following examples. A Belgian company vectorised maps of the Belgian NMA and used the geometric data thus acquired for the production of a digital road atlas of Europe. The NMA did not (and probably still does not) agree with this practice since it is of the opinion that the firm needs permission to reproduce the maps and should pay fees. Eventually, the NMA took the company to court, arguing that the use of the NMA’s maps by the company constituted a breach of unfair competition law. The Court of Appeal ruled that it was

LEGAL PROTECTION OF GEOGRAPHIC INFORMATION

121

not (Court of Appeal, Ghent, 1995). Although the firm profited from the maps of the NMA, it committed no act of unfair competition. The court came to this conclusion since:

¡ the maps used were only a minor source of data for the firm; ¡ the firm invested a lot of time and money in the vectorisation, and vectorisation is not ¡

the same as pure copying;5 the firm did not use the data to make a product that competes with that of the NMA, but to make a new product.

The German national postal services were more successful in their attempt to stop a competitor from copying (part of) their product. The German postal services publish a three-part telephone guide on CD-ROM with some 30 million records, each containing name, address and telephone number. The company buys these records at DM 2.90 each from its mother company. The CD-ROM is priced DM 3,950. A competitor advertised a CD-ROM with telephone numbers for DM 99, which turned out to contain a large number of records identical to those on the postal services CD-ROM. To produce their CD-ROM the competitor had also copied the computer program contained in the postal services CD-ROM. The court expressly stated that appropriating someone else’s achievement is in itself not an act of unfair competition, but since in this case the competitor took away the postal services’ competitive advantage by offering a competitive CD-ROM (at a fraction of the price), there was a breach of unfair competition law (Court of Appeal, Frankfurt, 1995).

Limitations on exclusive rights Three types of limitation on exclusive rights are discussed here. First, copyright acts themselves contain limitations. Secondly, laws on access to government held information might restrict the use of (exclusive) rights in information. Thirdly, the exertion of copyright is restricted by competition laws.

Limitations within copyright acts Every copyright act has provisions that allow for the free use of works in legal proceedings, for private study, etc. These so-called exemptions differ from country to country (Hugenholtz and Visser, 1995). Another restriction in copyright acts is that not all government information is subject to copyright, or only if certain conditions are met. In all EU countries except Ireland and the UK, official texts of the government and other public authorities, such as laws, decrees, judicial and administrative decisions, are not subject to copyright. The German act also stipulates that information published by the government in the public interest is not subject to copyright. According to the Dutch act, public authorities must expressly reserve their copyright on works, otherwise anyone is free to copy it or exploit it. Apart from the exemptions mentioned above, other works of public bodies throughout Europe may be copyrighted or fall under the scope of neighbouring rights, provided that they meet the relevant criteria for protection. The position of the public sector as producer of geographic information is, in this respect, equal to that of the private sector.

122

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Public access to government-held information National laws Access to government-held information is not directly related to copyright and other exclusive rights, but the issue has been looked into since public access might limit the exclusive exploitation rights on government-held geographic information. Not all countries have general public access acts. In those that do, it is sometimes stipulated that although persons have the right to obtain a copy of certain government documents (texts, maps, etc., either in paper or in electronic format), the copyright in them must be respected. In other countries this is not explicitly stated but held to be true none the less. Thus, it appears that general access acts only limit the powers of copyright holders to the extent that copies must be issued.

Directive on access to environmental information On 7 June 1990 a directive on access to environmental information was issued which obliges all EU member states to guarantee civilians access to environmental information (CEC, 1990). Environmental information is a broad term which covers information in any form or on any medium (text, sound or images, digital or analogue, etc.) regarding environmental regulations and policies, information on the condition of the soil, water, flora and fauna, etc. The information requested may be refused on a number of grounds, such as public security, confidentiality, etc., or because the information has been supplied to the government by a third party while the latter had no legal obligation to do so. Requests for information do not have to be motivated. Access to, and commercialisation of government-held information is an issue that is debated intensely at the EU level. Public access to government information is an issue that is raised in the EC’s greenpaper ‘Living and Working in the Information Society: people first’ of 22 July 1996. The European Commission states in this greenpaper that access to public information (services) is of great importance in improving democracy and social justice in Europe. The practicalities of broad public access, however, are not addressed. These include: what role should the private sector play in disclosing government information? does access for all citizens imply that information should be accessible (almost) free of charge? what are the implications of copyright in public sector information for access? No doubt these and other questions will be answered as public access policies are developed further.

Competition law Articles 85 and 86 of the Treaty of Rome (the founding treaty of the European Economic Community) contain important rules on competition, banning practices that are restrictive to free trade and the abuse of dominant positions. National laws by and large have similar provisions on trade within national territories. Therefore only the situation at the EU level is discussed here. For years, two joint cases were pending at the courts of the EU concerning information monopolies (based on copyright) and EU rules on unfair competition (European Union Court of Justice of the European Communities, 1995). These cases are important for the geographic information sector since they give an indication of the kind of behaviour that organisations with a dominant position in an information market are allowed. Quite a

LEGAL PROTECTION OF GEOGRAPHIC INFORMATION

123

number of organisations that deal with geographic information have a dominant position in a certain market, especially government agencies such as the census bureaux and the national mapping agencies. The EC Court of Appeal ruled on 6 April 1995 in what has become known as the Magill case. In connection with articles 30 and 59 of the Treaty of Rome (free circulation of goods and services), article 86 of the Treaty bans abuse of a dominant position in the European market. Publisher Magill argued that British and Irish broadcasting companies abused their dominant positions in the television guide market. Magill wanted to publish a complete television guide. Such a guide did not exist yet: viewers in the UK had to buy several television guides to get a comprehensive listing of all programmes. However, British and Irish public broadcasting corporations refused Magill permission to publish their copyrighted programme information, and in this way prevented competition for their own programme guides. It was, however, their regular practice to supply newspapers and other journals at home and abroad with excerpts of their programme information. Although the use of copyright powers in itself is not illicit in the sense of article 86, in special circumstances the use of copyright powers can constitute abuse of a dominant position. The Court ruled that this was the case, since the broadcasters used their copyright to prevent a new product (a comprehensive television guide) appearing on the market, while there was apparent demand. In addition, the broadcasters did not have any intention of producing this new product themselves. The EC courts only rule on cases where the European market is affected, that is to say where the free circulation of goods or services among two or more member states is involved. For producers of geographic information the Magill ruling means, for instance, that if producers of different countries join forces in some part of the geographic information market and this results in a dominant position, they will not be allowed to block the production of value added products by others if they do not (intend to) produce these products themselves. The court ruling mentioned here, together with others, left its mark in earlier drafts of the database directive. The drafts contained provisions for compulsory licensing schemes, in cases where producers have a natural or legal monopoly on information. These provisions were eventually dropped (see above) but in the preamble to the directive it is stated that rights in databases may not be exercised in such a way that it results in the abuse of a dominant position. Thus, the provisions of the directive are without prejudice to the application of national or European rules of competition (of which the Magill case is a good example).

Conclusion The ultimate goal of EUROGF’s activities in the area of legal protection of geographic information is to encourage the development of a legal framework that is stimulating rather than a hindrance to the development of the European geographic information community. A first step has been taken by studying the current situation concerning legal protection of geographic information in the member states of the EU. For an adequate legal framework to exist, first and foremost transparency is required with regard to what type of geographic information is subject to (exclusive) rights, to what these rights pertain and who is the beneficiary of the rights. No less essential is that the legitimate interests of both users and producers be balanced, as well as the interests of both the private and public sectors.

124

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Type of protection On surveying the type of legal protection available for geographic information in EU member states, one notices that there is no such thing as a transparent system. European copyright laws are alike with regard to the types of work that qualify for copyright protection, typical products of geographic information such as maps often being mentioned explicitly as subjects of copyright. The all-important requirement of originality is interpreted similarly in European countries, with Ireland and the UK being notable exceptions. Although collections, whether of items of information or data, are possible objects of copyright in any of the countries researched, the required originality may prove to be prohibitive (except in Ireland and the UK). With regard to the protection of non-original information products, the situation is very diverse. The Nordic countries protect all producers of collections of data/information through the catalogue rule. So does Dutch law, in a more restricted manner. Other countries do not have similar general provisions. Sweden, Portugal and Germany, however, have legislation that specifically protects certain producers of geographic information, notably the national mapping agencies. Non-original photographs are also protected in a number of countries, including Austria, Denmark, Finland, Italy, Spain and Sweden. Another type of protection is that of publishers (Greece, Ireland and, to a limited extent, Spain). In addition to the diversity in protection through national copyright and related rights, the extent to which producers can invoke unfair competition law to protect their interests against competitors differs from country to country. The upcoming implementation of the EU directive on the legal protection of databases will change the situation considerably. Not only does the directive ascertain that databases are protected by copyright if they are a personal intellectual creation of the author (i.e. original), but it also envisages a new right concerning databases that are not eligible for copyright protection, namely the right of producers to prevent extraction of all or a substantial part of the data from their databases. The impact of the directive on the development of the geographic information sector is at this point difficult to assess. No doubt the directive greatly improves protection for producers of geographic information. Important questions remain, however. What, for instance, is a substantial investment in a database warranting protection? How likely is it that producers of frequently updated databases will enjoy everlasting protection, possibly stifling the production of valueadded products based on these databases? Will it in practice be possible to make a clear distinction between computer programs incorporated in the database (not protected under the directive) and the collections of data which are subject to the right to prevent unfair extraction? Does the directive meet the interests of users well enough?

Ownership of rights Since geographic information may be subject to different types of right, the identification of what person/company has which rights may prove to be difficult and time consuming to carry out. In cases of cross-border cooperation, with input (maps, data) coming from different sources, the issue of ownership is particularly complex. Rights owners, for instance, may be a mix of the following: the natural person or persons as author(s) in the sense of copyright; the producer of a database in the case of catalogue protection or, in the near

LEGAL PROTECTION OF GEOGRAPHIC INFORMATION

125

future, under the database directive; the photographer in the case of a photograph; the employer in the case of works created by employees. In addition, exploitation rights may be transferred, or given in licence. Although, on the one hand, transfer or licensing of rights increases the difficulty of ascertaining who owns rights, it gives, on the other hand, the possibility of regulating by contract the powers of the parties involved. Good contracts will remain important even though, in the future, electronic rights management systems might be developed.

Limitations on exclusive rights The situation in EU countries with regard to restrictions on rights in information is diverse, particularly with regard to government-held information. In some countries all works of the government are subject to exclusive rights (UK, Ireland), in others the copyright laws restrict the possibilities of public authorities to own copyright or exercise copyright powers (Germany, Netherlands). With regard to the relationship between rights of the public to access information and access to government information, differences also exist. Some countries do not have general access laws, but in those that do, the fact that copyright or otherwise protected information can be obtained by invoking public access acts does not imply that the information received may be exploited commercially. At the European level there is concern for improving access to public sector information, but there is not yet a clear set of ideas on how this should be attained in practice.

Where to go from here As for the activities of EUROGI, a follow-up study has been instigated with a view to working on solutions to the problems concerning legal protection of geographic information. In addition, EUROGI has reviewed which other issues (to name a few: liability for information services and products, privacy, access to geographic information, commercialisation of geographic information held by the public sector) need to be addressed and in what order.

NOTES 1. In legal theory a difference is made between copyright and droit d’auteur. The former prevails in common law countries, the latter traditionally in continental Europe. In this chapter the term ‘copyright’ is used to indicate both systems. 2. Exemptions that allow the free use of copyright works for private purposes, in legal proceedings, etc. have not been studied. For an analysis of this issue see Hugenholtz and Visser, 1995. 3. The difference between transfer and licence being that in the latter case the owner retains copyright while giving others permission to exploit the work. Moral rights, such as the right to be named as author, to resist changes being made in the work, etc. may not be transferred, but can be waived under certain conditions. 4. The acts on unfair competition (also called Marketing Act, etc.) contain provisions on various subjects, such as illegitimate pricing strategies and the (illegitimate) imitation of products that causes confusion among the public as to the origin. 5. Note that ‘copying’ here is done through vectorisation rather than rasterisation. Digitising by rasterisation might be considered to be more of a direct type of copying.

126

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

REFERENCES EECHOUD, M.M.M.VAN (1996). Legal Protection of Geographical Information. Amersfoort: EUROGI. CEC (1990). Directive 90/313/EEC, Access to Environmental Information, 7 June 1990, OJ L 158, 23 June 1990. CEC (1991). Directive 91/250/EEC, The Legal Protection of Computer Programmes, 14 May 1991. OJ L 122, 17 May 1991. CEC (1992). Directive 92/100/EEC, Rental and Lending Rights and on Certain Rights related to Copyright in the Field of Intellectual Property, 19 November 1992, OJ L 346, 27 November 1992. CEC (1993). Directive 93/98/EEC, Harmonizing the Term Protection of Copyright and Certain Related Rights, 29 October 1993, OJ L 290, December 1993. CEC (1996). Directive 96/9/EEC, Legal Protection of Databases, 11 March 1996, OJ L 77, 27 March 1996. HUGENHOLTZ, P.B. and VISSER, D.J.G. (1995). Copyright Problems of Electronic Document Delivery: A Comparative Study. Brussels: CEC DGXIII.

Statutes and legal cases Berne Convention for the Protection of Literary and Artistic Works, 9 September 1886. Court of Appeal, Ghent (1995). (Hof van Beroep te Gent 7e kamer), case no. 1994/A.R./0853, Nationaal Geografisch Instituut v.Tele atlas NV, 16 November 1995. (Unpublished) Court of Appeal, Frankfurt (1996). (Oberlandesgericht Frankfurt), case no. 11 U 10/95 4 July 1995, in Computer und Recht 4, 112–15. Court of Justice of the European Communities (1995). Case numbers C241/P and C242/P, 6 April 1995. Universal Copyright Convention, 6 September 1952 as revised at Paris on 24 July 1971.

CHAPTER TEN

Intellectual property and mapping: a European perspective LAILA ASLESEN

Introduction GIS is moving from electronic versions of traditional mapping and charting products into multimedia systems; where the products create the basis for the use of other types of work. An endless possibility for new applications arises. A number of legal aspects also arise, many just as important as intellectual property. This chapter gives an overview of the legal situation on intellectual property and GIS, and focuses on some specific GIS problems, as well as the dilemma of how to make data available without compromising protection.

Legal protection Introduction Copyright is a term for the set of rules that protect authors of literary, artistic or scientific works. It originated from unfair competition principles, where no-one could unfairly exploit another’s work. From the start it was a protection for those who invested money in works. It was gradually accepted that copyright protection allowed publishers and other investors unfair exploitation of a creator’s work. It was necessary to have rules protecting creators. Investment would be protected by unfair competition and contract law (contract with the author and unfair competition law to protect against any other third party). Neighbouring and complementary legal regimes have also been established. It is important to remember that selling intellectual property is different from selling physical property. The important thing is not the customer’s use of a book or a disk, which is only a way of distributing the product. What is important is what the customer is allowed to do with what is on the disk or in the book: that is the real product.

127

128

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Copyright protection The Berne Convention As explained in the previous chapter, the prevailing international regulation on copyright protection is the Berne Convention (BC).1 The BC has minimum requirements for national laws within the Berne Union (nations signing the BC are regarded as a union). The BC uses the principle of national protection.2 Foreign works have the same protection as national works. It should be remembered that works can originate in more than one country depending on the citizenship of the author and the place(s) of publication. Protected works are classified as being from the literary, scientific and artistic domain. The examples mentioned in the BC are not exhaustive. Works are protected in whatever form: digital, analogue, or any other form. The BC gives no further requirements for protection, but it is commonly held that copyright protection does not apply to all works. There are minimum requirements (varying from country to country) which can be described as ‘level of achievement’.

Level of achievement Though not explicitly stated in the BC, there is a bottom line for what can be protected even if it is a literary, scientific or artistic work. One way of expressing this is to say that facts fall below the level required to gain copyright protection unless they are expressed in an original manner. Original means here that it is not a copy of someone else’s work, but the result of the author’s own labour. It will be the expression that is protected, not the facts. In most countries, originality would not be sufficient to gain copyright protection. It is also necessary to demonstrate that the work is an expression of an individualistic intellectual activity on the part of the author. This intellectual activity may be a result of technical skill and experience, and does not necessarily have any artistic or aesthetic element. Literary, scientific or artistic value is of no importance. The author can use ideas and thoughts that were proclaimed by others. The tools that are used are not significant. One test as to whether the author has put an individual mark on the work is to determine if the work is of a type that can have many modes of realisation. If the work, by its nature, has very narrow limits for how the result can be formed, there will often be no place for individual intellectual activity.

Author’s exclusive rights and their limitations The BC gives certain exclusive rights to the author of a work.3 It allows for exceptions from these exclusive rights, the extent and details of which are left to national legislation.4 National regulations have to be interpreted in such a way that they do not ‘conflict with a normal exploitation of the work and [do] not unreasonably prejudice the legitimate interest of the author’. The BC regulates the relationship between author and adapter.5 Adapted works cannot be used without the consent of the original author. When they are used legally, the adaptations are protected against misuse from both the original author and others. This also applies if the original author does not enforce his or her copyright: an important point for users of works of US government origin.

INTELLECTUAL PROPERTY AND MAPPING

129

Other types of protection In some countries where originality is not sufficient to give copyright protection, works with originality but no intellectual achievement are given a separate type of protection. Telephone directories are an example of such a work. The EU Directive 96/9/EC of 11 March 1996 on the legal protection of databases, introduces such a type of protection for analogue or digital databases that are not copyright protected.

Protection for geographic information products Copyright in the product as such Geographic information may be expressed as a map or chart, a table, a graphic presentation, a coordinate, or in some other form. In each case one must consider whether the product in question fulfils the requirements of copyright protection. Charts and maps are not mere compilations of facts. They include an extensive generalisation of facts (the smaller the scale, the more the generalisation). Furthermore, they are presented in a way that is carefully designed according to the type of map, the scale and general readability. Maps and charts are as such copyright protected. The interesting point is to what extent are the elements of a map or chart protected? In legal theory not only is the work itself protected, but also copying of parts of the work can be considered an infringement. In common law systems one talks about substantiality. If a substantial part of the work is copied, this is an infringement. Substantiality is not only measured in quantity but also in quality. In other legal systems one talks about whether the elements in themselves are original enough to be copyright protected. The general idea is the same: you cannot circumvent copyright protection by copying parts of the work.

Copyright in the elements of the maps and charts It is the generalisation and expression of facts that are protected. Facts read from a map may be copied, but not used to make a map that is formed on the basis of the first map. If the facts and expression are merged in a way that makes the fact accessible only through copying the map, the courts may say that this is a copyright protection of facts. An example is found in the US Doctrine of ‘Merger’.6 In other countries this could be considered as a lack of originality and intellectual achievement, rather than a copyright protection of facts. The author of the map will have to decide which geographic names to use, where to place them and which fonts, sizes and colours to use. This is not done by chance, but carefully chosen to produce the best cartographic result. The form and selection of geographic names is therefore protected, and even more so as a part of a total cartographic expression. Depths and heights are, in essence, facts. When included in a map or a chart they may, however, be representations of facts as well. The height of a mountain will usually represent exactly how high the mountain is. A depth on a chart might, however, not be the exact depth at that point, rather a generalisation that indicates that it is at least that deep on and around it. The depth has been selected by the hydrographer, in conjunction with others, as a generalisation that gives maximum navigational information without making the chart too full of information. It is therefore doubtful whether one can copy information on depths and heights and claim that this does not conflict with the copyright.

130

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

By producing contour lines the cartographer makes a generalisation of height and depth observations. A 20 metre contour line will tell the mariner that the depth here is at least 20 metres, but usually very few points on the line will be exactly 20 metres deep. It may be 21, 22, or even 30 if there is a sudden cavity that was too small to be taken into account in a chart of this scale. This does not mean that the chart is inferior, only that a generalisation has to be made. A chart or a map cannot give a complete picture of reality unless it is made in scale 1:1. A contour line is one of the most skilled and difficult parts of chart- and map-making and will have the necessary level of achievement to gain copyright protection. An interesting issue is the use of software to generate contour lines. There is no software tool today that can make contour lines ready for use from the raw data. Even if such software existed, one would have to establish that the product in question was made solely by using this software, if it should have any influence on the copyright protection.

Problems related to GIS It is difficult to sort out the legal context of putting together a GIS. If somebody wants to use a literary, scientific or artistic work, they would have to look to the national legislation. Origin and ownership of the work must be determined. Different versions of the BC exist. Several forms of protection may be relevant. Finally, most of the problems raised by modern technology are merely discussed among scholars and bureaucrats at this point. Legislation from the analogue age is, to a certain point, easier to use than what has been made in the past ten years. Regulations have been made dependent on the new technology and have to be changed along with it. This means that legislation is always a step or two behind the new technology. However, legislation should be independent of technology. It is important that producers of geographic information find a system for conveying permissions to use their data that can remedy the shortcomings of legislation. The system has to be easy to fathom and cost efficient.

Legal protection The reason why protection of each element in a map is important is that digital technology makes it possible to separate elements and re-use them in a very efficient way. It is therefore important that the elements that represent the real creative work are protected. It is equally important that the GIS-makers know to what extent their use requires someone’s permission. In this context the new technology raises several issues.

Reproduction Each analogue copy, on paper and similar material, is a reproduction according to copyright law. Copying from one electronic storage device to another, for example, from hard disk to disk to CD-ROM, is also a reproduction. This is a problem given usage in a digital world. It will mean that the copyright owner virtually must give permission for any kind of practical use in a digital environment. The sensible approach is to use the old analogue principles as a guideline. These principles are connected with the legal ideas of use versus reproduction and are not technology-dependent. A very good principle is embodied in the BC: reproduction is

INTELLECTUAL PROPERTY AND MAPPING

131

within an accepted and legal use if it does not ‘conflict with a normal exploitation of the work and does not unreasonably prejudice the legitimate interest of the author’. If it falls outside this it should be a reproduction that requires permission. Reproduction within a PC does not conflict with normal exploitation. As long as the use of the digital copies stays with the person or company that bought the original, the same applies. Screen displays In Norwegian law a copy is any sufficiently lasting affixation of the work. A screen display is not sufficiently lasting. If, however, the work is converted to adapt it to the screen display, this adaptation may need the author’s consent. That is, if adaptation involves making a sufficiently lasting affixation of the work within the hardware, then the author’s consent must be sought. This is probably the way the conversion will be argued in most European countries, with the exception of the UK, where screen displays are regarded as a copy. It is likely that new solutions will arise in national laws and perhaps in EU legislation. Other theories have been raised as to how screen displays could be regarded. One point of view is that the screen display is a way of communicating the work by using an instrument that transmits the work by images.7 In that case it is only public communication that is within the author’s exclusive rights. Another point of view is that a screen display is a work in its own right and not merely a display of the underlying work. Adaptations and new works The border between adaptations and new works is difficult to decide on a general basis. If the adaptation has a sufficiently high level of achievement then it is a new work. In a digital environment, adaptations are often easy to accomplish with a good result, which makes this border increasingly important. As long as the adaptation uses a substantial part of the copyright-protected work, it is not a new work. A good example is a navigational chart. When a chart is produced there is a careful selection and adaptation made with safe navigation in mind. It is important that the representation of coastlines, contour lines, depths and navigational aids enhances the work of the navigator. If someone copies it, still for navigational purposes, it is not enough to make further selections, to change colours, lines, fonts or scale. It would still be a copy of the careful and skilled work that is the real value of the chart. The fact that the result of the work looks quite different from the original cannot be decisive. To obtain a different look does not necessarily require much intellectual input in a digital environment. Physical protection There are several ways of stopping unwanted use of or access to data, like fingerprinting, hardware locks (dongles) and encryption. Physical protection can, of course, be circumvented. It is important that legislation makes circumvention of such physical means illegal. Administration and pricing The Norwegian Mapping Authority (NMA) uses a system for selling rights as a basis for pricing policy and administration. The starting point is the protection provided by Act no.

132

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

2 of 12 March 1961 relating to copyright on intellectual property (Copyright Act), which incorporates catalogue protection and other neighbouring rights (broadcasting, performance rights, etc.) The following discussion on copyright includes catalogue rights. The NMA defines itself as the copyright owner, with the exclusive right to exploit its works. These rights can be transferred in whole or in part to others. The most important of these are the rights to reproduce. Reproduction can be done with the copyright owner’s consent or according to regulations given under the Copyright Act. In Norwegian law these exemptions originate from the theory called ‘rights of the general public’. The public has an independent right to use works based on its need for cultural input and scientific information. Without a definition of what one sells it is difficult to set the right prices, and hopeless for the customer to understand what he receives for his money. The NMA has defined certain rights that it sells. The definitions take into account the rights of the public and cover usage rights and marketing rights.

Usage rights The right to use the product for internal or private purposes is a usage right. If the customer is a private individual, copying and presentation are permitted for private use. Production of copies by transferring analogue products to a machine-readable medium (digitising/scanning) is not permitted without special agreement. This right is called the right to digitise (this will go out of use gradually as customers are able to buy NMA’s digital data). The reason for this is that it could conflict with the NMA’s exploitation of its works. Usage rights are the basic concept of the NMA’s pricing policy. These rights are what customers acquire when they buy an NMA product. If they want to add value or otherwise rework the product for internal use, they can do that under usage rights. If they start to distribute copies on a larger scale, they will be considered as being outside usage rights and into marketing rights. Usage rights do not include the right to copy for internal purposes in a company or similar legal persons. Previously the NMA allowed such copying to take place because it did not see it as practical or even desirable to monitor such copying. The Copyright Act has now been amended to make it possible for all kinds of legal persons to acquire a collective licence agreement on such copying, and it is preferred that the NMA’s customers do that instead.

Marketing rights These are rights to produce copies of the product, or an adaptation thereof, and to sell or otherwise distribute these copies. The customer must acquire usage rights before he can acquire marketing rights. Many of the NMA’s standard products are sold through distributors. The distributors do not have usage rights to the product. Distributors of digital products may adapt the product to their hardware or software, if this is necessary for their distribution. For each of these rights the customer pays a fee. The usage fee is the normal sales price of the product. For marketing rights, the customer pays a marketing fee calculated on the basis of three elements: costs, loss of revenue and possible mark-up.

INTELLECTUAL PROPERTY AND MAPPING

133

Availability versus protection Authors, owners, makers and users GIS is a fairly new concept, which raises lots of problems concerning the protection of authors and owners on one side, and the interests in availability and low prices on the other. There is a growing need to develop copyright law and find out how it can be applied in today’s society. For the government agencies that produce maps and charts, the reasons for this growing need are twofold:

¡ governments are cutting costs and looking for possible sources of income; and ¡ the technological developments make what was a more or less satisfactory protection in the analogue world unsatisfactory in a digital age. This is not an ideal situation for either the agencies or the users of their products. There has been an outcry for ‘free public information’, often without a clear definition of what ‘public information’ or ‘free’ are supposed to be. It is probably not practical politics for any European government to pay for establishing geographical information databases without any cost recovery from the sale.

Reasons for protection Intellectual property protection is the guarantee that makes it possible to invest in the quality information needed for GIS. It is, however, important that the producers do not price or sell the products in a way that makes it difficult or impossible for interested GIS users to avail themselves of the best possible product for their individual use. On the other hand, if the protection is insufficient, it will influence both the will to make good products and the will to put them into circulation in a way that benefits the market. One could get a situation where the market is polarised into one, easily available market with low quality data, and one market for those with the money to buy high quality data. This may not be a bad thing, but lack of money does not always mean that one can be satisfied with low quality. Many government institutions and charitable organisations are engaged in tasks that are vital to us all, but their funds can be very limited. Low quality and simplicity are not the same. For many GIS purposes, detailed, accurate data are needed. For others, less detail is required but they should still be as accurate as possible under the circumstances. The simpler versions should be cheaper, but to simplify data might actually demand a lot of extra work for the producer if the result is going to be acceptable.

Pricing and value The problem facing GIS makers and users is not that the information they want to use is protected, but that protection means expense. It affects their administration as well as requiring payments to the owners and authors. It is tempting to say that data must be as cheap as possible to avail ourselves of the new technology. There is, however, a danger that if we focus on cheap data, users will not understand that production of high quality geographic information is an expensive process.

134

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

In an information society, the value is in the information, not in the means of distribution and use. Some argue that if a customer pays a certain amount for the hardware and software of a GIS, it is not acceptable to pay more than so much for the information to use in it. The question is, what is he going to do with the hardware and software without any information? If we tell the customers that the quality is the reason that the price is so high, they need an assurance that when they pay this price they will receive quality. For this we need international standards and procedures to control and guarantee the quality level. Creating a database can be a big and costly job, but in many cases the producer can then rest and gather income. For geographic information the situation is different. If the information is not updated regularly, the value of the data drops dramatically after a year or two. Updating is an important aspect which many tend to forget. It raises legal questions as to the status of works made by one party and updated by another. To establish geographic information databases it is necessary to have cooperation between government agencies, local communities, international organisations and private companies—and this is vital to all of them.

Administration An important challenge is the process of identifying what the GIS maker wants to use in his product, how and where to buy it, and considering the pricing implications of his product. Basic map and chart information is for the most part produced on a national basis, but GIS will often be international in its scope, as can be seen from the example of charts. To get permission to digitise a chart for a seamless electronic navigational database, private companies naturally go to the chart publisher, like a national hydrographic service (HO). These charts may, however, contain protected information from other HOs, obtained by the publisher according to agreements within the International Hydrographie Organisation. According to these regulations, the HO in question would have to refer the company to the other HOs to get permission from each and all of them for every single chart they wanted to digitise. This is a costly and ineffective process. Certain HOs have therefore started to develop a system called custodianship, which means that the chart publisher can give permission for all the charts they publish and channel the income back to the others.

Conclusion The challenge of the GIS business is to create the necessary legal tools to protect its products, so that it can trust the information distribution systems and create the necessary standards and quality management to give people easy access and trust in what they get. It is not possible to build an information society without adequate intellectual property protection. Any solutions will, however, have to give due consideration to other important legal aspects, such as liability, competition law and protection of privacy. Legal rules are the tool with which to build a society: a good tool only if used wisely and with foresight. The GIS business will have to help the lawmakers in developing the tools that are needed for its use. At this stage the lawmakers are still looking at the problems in most countries. If the business itself develops practical and sensible solutions, the lawmakers will have a much easier task, and the results will be good for everyone.

INTELLECTUAL PROPERTY AND MAPPING

135

From the basis of intellectual property law, one should go on and look at what other measures can be taken. Given the current state of legislation in this field, intellectual property law will have to be the last resort. There are two markets that have to be treated differently: the consumer mass market and the market for professional and research users. In the first case, physical protection is the obvious solution to try. In the other market, contract law can be a useful tool. In these matters authors, owners, GIS makers and users have common interests. They should work together on an international as well as a national level.

NOTES 1. This chapter uses the text after revision in Paris, 1971. 2. Article 5.1: ‘Authors shall enjoy, in respect of works for which they are protected under this Convention, in countries of the Union other than the country of origin, the rights which their respective laws do now or may hereafter grant to their nationals, as well as the rights specially granted by this Convention.’ 3. Article 9.1: ‘Authors of literary and artistic works protected by this Convention shall have the exclusive right of authorising the reproduction of these works, in any manner or form.’ Article 11 bis. 1: ‘Authors of literary and artistic works shall enjoy the exclusive right of authorising…(iii) the public communication by loudspeaker or any other analogous instrument transmitting, by signs, sounds or images, the broadcast of the work.’ Article 12: ‘Authors of literary or artistic works shall enjoy the exclusive right of authorising adaptations, arrangements and other alterations of their work.’ 4. Article 9.2: ‘It shall be a matter for legislation in the countries of the Union to permit the reproduction of such works in certain special cases, provided that such reproduction does not conflict with a normal exploitation of the work and does not unreasonably prejudice the legitimate interest of the author.’ Note also the exemptions in Articles 2.8, 10.1 and 10.2. 5. Article 2.3: ‘Translations, adaptations, arrangement of music and other alterations of a literary or artistic work shall be protected as original works without prejudice to the copyright in the original work.’ 6. ‘It may be permissible to copy, the selection and arrangements of facts, for example, without incurring liability for infringement due to merger. That is, under the merger doctrine, if ideas or facts can be expressed in only a limited number of ways, then even the expression, which is normally protected by copyright, may be copied; otherwise the unprotectable facts or ideas would be effectively monopolised by the first person to express them. So in the case where a court determined that a map of the proposed location of a pipeline was the only way to depict the idea of the pipelines location, the map was non-copyrightable because there was no other way to depict the idea of the pipeline, Kern River Gas Transmission Corp. vs. Coastal Corporation, 889 F.2d 1458 (5th Cir.), cert, den., 498 U.S. 952 (1990).’ (Samuel, ‘Navigating the law of copyright: an overview of copyright law for those who make and use electronic navigational charts’, 2nd International Conference on Maritime Law and the Electronic Chart). 7. BC Article 11 bis. 1.

CHAPTER ELEVEN

Legal and institutional issues to be resolved with respect to the integration of European data ANDREW LARNER

Introduction The integration of geographic data raises a number of legal and institutional issues. The importance of particular issues will vary among member states. This chapter is written from the UK perspective. In some ways the UK position is an anomaly within the EU because of the emphasis on government cost recovery and cross-charging within and between public sector organisations. However, it is contended that the UK market approach has reflections in other member states and that these similarities may become stronger as the financial effects of the increasing average age of the European population are felt. Whilst the position taken in this chapter is based upon the reality of charging for government data, it is not intended to argue for or against government cost recovery. The purpose of this chapter is to try to describe, given the realities of the data market, what is required to ensure an equitable balance of interests in data. It is contended that this balance is the fundamental issue. Before considering downstream issues such as liability for the quality of data, it is first necessary to clarify the nature of data, the parties who should have rights in data and to define those rights. In addition to the existence of rights, the way in which the legal framework administers those rights and adjudicates between the parties exercising them may affect the ability to integrate data sets. It is this latter aspect of the legal framework which is referred to as ‘institutional restrictions’. To be of significance in the context of the EU, the issues associated with establishment and adjudication of rights in data must relate the establishment of an unrestricted European market. It is not the purpose of this chapter to justify the harmonisation of rights in data within the EU. However, it is worth noting in passing that such a justification has three components: transborder, including pan-European, products; pan-European companies; and mobility of labour within Europe. There is a relationship between a European market and rights in geographic data for transborder products and products produced by panEuropean companies. This relationship calls for a harmonised system of rights within the EU. In addition, we must consider the mobility of professionals in the European geographic information market. A harmonised approach to the professional management of rights in data would allow greater freedom in the European labour market. 137

138

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Table 11.1 Areas of the law which are of concern

Background To consider a legal framework for rights in data one needs to examine a wide variety of areas of the law (Hammond, 1986, p. 344). Hammond states that, ‘It is difficult to “carve out” a piece of the whole in isolation from the “system” that the whole is supposed to support’. However, if a common understanding of the professional handling of geographic data is to arise then the legal framework for geographic data must be defined. There is ample evidence within the GI community of attempts to define the areas of law of concern. These areas may be identified in two ways: areas of the law which are thought to be of concern, and aspects of handling geographic data which are thought not to be handled adequately by the existing legal framework. Table 11.1 identifies the areas of the law that have been associated with the certainty of rights in data by a variety of parties from the legal and GI communities. The Legal Advisory Board conference minutes include a variety of perspectives including that of the data broker (Legal Advisory Board, 1994). The data broker may be viewed as a purchaser, user and value added creator of data. However, where European law is concerned it is likely that the majority of opinions represented will come from the vendors’ perspective rather than the users’. Having identified areas of the law of concern, the tendency is to do one of two things:

¡ describe how the law works in general; ¡ describe what UK users must do to use and/or comply with the law. Undoubtedly these reactions are valuable. General education about the way in which the law works is essential for users of information, as is guidance about how to avoid falling foul of the law. However, to limit consideration of the law to these reactions is to deny the possibility that new circumstances have arisen that make the current legal framework

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

139

Table 11.2 Detailed concerns of the UK Gl community

inappropriate for the interests of society. In order to consider how the law ought to work, it is necessary to know more detailed concerns of those affected by the handling of UK law (see Table 11.1). In the UK the Association for Geographic Information (AGI), the umbrella organisation for the GI community, has facilitated a number of discussions that reveal the concerns of the UK GI community. The AGI published two landmark reports, one in 1992 and the other 1993, that represented the results of significant consultation exercises. The first report considered the role of copyright in the GI community (AGI, 1992) whilst the second considered the pricing policies of the Ordnance Survey (OS) of Great Britain (AGI, 1993). Both of these reports were born out of user concerns regarding OS copyright licences for their digital data. The AGI reports were very useful in dispelling misconceptions about the intentions of Ordnance Survey and set the foundation for positive discussions between users and other major GI providers. Whilst not clearly evident in the published documents, the investigations that underpinned the AGI reports were very useful in that they clearly identified what the GI community wanted from a legal framework for data. Table 11.2 shows the concerns expressed by those giving evidence to the AGI (1992) regarding the application of the law

140

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

to the handling of digital geographic data. The opinions are categorised according to the background of the organisation expressing the concern. The most universal concern expressed relates to the handling of value added data. ‘Value added’ data creation occurs where a party takes existing data and creates a new data set. The processes used may merely combine existing data (‘merge’) or they may calculate the values of a new data set (‘derive’) by applying rules to the existing data. It can be seen from Tables 11.1 and 11.2 that third party rights to data must be considered. In Table 11.1 data protection, privacy and access to public sector data have arisen as issues of concern. Table 11.2 also lists access to public sector data as a concern.

Ownership rights in records of facts If the primary concern is to establish the rights in a data product such as its ownership, who may use it and who may restrict its use, it is also necessary to consider the balance between such rights. However, before the rights of owners, users and third parties can be determined, it is necessary to have an unambiguous definition of data. Currently no universally accepted definition of this term exists. Furthermore, no legal definition of data exists. The definitions used here are those established in Lamer (1996). A fact is a property of a real world object. A data set is a compilation of records of facts. Explicit records of facts may be referred to as data items. For a data product to be owned with certainty it is necessary to own the records of fact within that product. However, the law seems to consider that the fact and the record of the fact are synonymous. Currently neither the fact nor the record of the fact are expressly protected by law. In copyright law the data set is protected as an original literary work by virtue of the selection and arrangement of those data contained within it. The contents of the data set are often incorrectly referred to as facts rather than records of facts. Furthermore, the law considers these facts to be the idea, and their selection and arrangement to be the expression of the idea. It is held that ideas, and thus by inference the contents of the data set, cannot and should not be owned. Consequently, the desire to balance rights in data is in conflict with the law’s desire to ensure the free flow of ideas. It has been submitted that a fact is a real world object separate from the record of the fact. Thus owning copyright in the record of the fact does not stop the free flow of ideas. Two parties may independently make records of the same facts and distribute their data sets as they see fit. Furthermore it is argued that the record of the fact is the expression of the idea, whilst the selection and arrangement of the such records is a part of the idea. The idea underlying the data set is the cognitive model of the real world. Thus in a new legal regime created for data, the creator of a record of a fact should own that record. Since data may not be owned, parties with interests in a data set must try to effect their ownership through protection of its selection and arrangement. For the selection and arrangement of a data set to be protected it must be original. As pointed out by van Eechoud (Chapter 9), UK copyright law has a low test of originality that merely requires that the work has been created by virtue of the sole effort of the author. In other EU member states the test of originality is higher. The difference between member states’ approaches to originality in the context of data sets may be classified as follows:

¡ the author has arranged the data items; ¡ the author has exercised choice in the selection and arrangement of the data items; ¡ the author has shown novelty in choosing the selection and arrangement of the data items.

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

141

On the basis of lack of originality, copyright protection has been denied to maps in many member states. In the UK the courts have held for many centuries that maps are protected by copyright. It might therefore be felt that the UK’s approach is more desirable. The link made between protection and effort has meant that the courts in the UK, when trying to detect copying, may merely look for the use of somebody else’s labour. When applied to protecting data sets, this approach sometimes protects records of facts as a by-product of protecting their selection and arrangement. In such cases the debate focuses on how many records of facts must be copied before the selection and arrangement are copied. In many other countries when the protection of the selection and arrangement of a data set would protect records of facts, then neither would be protected. The reason given for withholding the protection from the data set would be either that the idea and the expression had merged or that the expression was not sufficiently original. In the UK, copyright has always had an economic imperative and even where the courts hold that copyright does not exist, infringement of the copyright owner’s rights may occur and warrant damages (see, for example, Kenrick v.Lawrence [1890]). Data set vendors wish to protect their investment by ensuring that users’ handling of data does undermine further sales. Use of works within a user’s organisation does not, as such, deprive the data set vendor of the ability to make a profit. Consequently, any reproduction of the work incidental to its use should not be an infringement of the data set creator’s rights. Reproduction in this case includes, copying, adaptation or creation of an original work from another. Circumstances where sales are lost occur when a single copy of the work is copied onto more than one computer. Consequently, the licensing of terminals is of concern. The number of terminals licensed may be calculated according to the number of terminals that store the work or the number of terminals that may use the work concurrently. As noted previously, a data set contains explicit and implicit records of facts. By comparison of explicit records of facts the user may create new records of facts. These new records are considered to be implicit within the data set. The extraction of implicit records of facts requires intellectual effort on the part of the data set user. Consequently, records of facts held implicitly within a data set should have two authors: the data set creator and the data set user. Therefore derived works should be jointly owned by the data set user and the data set creator. The derived data set may be used as a summary of the source data set. Where this is the case and the derived data set is used on more machines than are licensed for the source data set, then the creator of the source data set may have a legitimate complaint. It is submitted that in the majority of cases the use of the derived data set within the user’s organisation will have little effect upon the ability of the creator of the source data set to profit from his or her work. Currently the default position is that no derived work may be created without the permission of the creator of the source work. The default position should be that users have unrestricted use of derived works within their organisations. However, the creator of the source data set will need easy access to adjudication in order that a loophole in the protection of records of facts does not occur. The effects of dealing in a derived work are different from simply using it inside the user organisation that created it. Even if the derived work is not a summary of the source work, a sale of the derived work would profit from the labour of the creator of the source work. If the derived work is a summary of the source work, then in addition to profiting from the source work, sales of the source work may be affected. If the balance of interests is with the creator of the source work then the creator of the source work has a monopoly

142

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

in derived products. This restricts the market and is a move towards owning the fact rather than the record of the fact. If the balance of interests is with the creator of the derived work then the incentive to create the source work is reduced. This is an erosion of the ability to own the records of facts. It is submitted that, as with co-ownership of real property, any joint author should have the right to realise his share of the property. Consequently, a joint author should have the right to sell copies of a derived work without permission of the other joint authors. However, the administration of the rights in data sets should allow for the settling of disputes over the value of a work and the division of shares in a work. Under the Berne Convention, current copyright systems give the creator of a data set economic and moral rights. Economic rights are alienable whilst the moral rights may be waived. The role of moral rights has been reduced with respect to digital works. The importance of ‘notice’ in copyright law suggests that the right to be identified as the owner of copyright is as important as being identified as the creator. Thus the right of association may be seen as facilitating the creation of copyright works. Conversely, the ability to object to derogatory treatment of the work may be seen as restrictive of creation, particularly with respect to derived works. It is submitted that with respect to data, moral rights should be reduced to the right of association. This right should be reserved for the owner of copyright. Whilst it has been submitted that the position of moral rights should change, the same does not necessarily follow for other aspects of copyright. It is unlikely that an individual will be the author of a data set. The corporate nature of the creation of the data set combined with the protection of records of facts requires a consideration of the first owner of copyright. Normally the creator of a work is the first owner of copyright. The economic nature of the right would suggest that the employee would not normally be the owner of copyright. Thus the individual would own her creations only when clearly working outside her terms of employment. The difference in practice is that employees are unlikely to be given permission to create data using data or equipment in their employer’s possession. Thus the distinction between works created as an employee and those created outside the contract of employment is likely to be clear. It is submitted that the position regarding contractors would also remain unchanged. The rights suggested above run with the data set rather than the parties to the rights in the data set. For example, the right to be identified as the copyright owner would arise from ownership and not require assignment. Consequently, it is logical to tie the creation of a right to the creation of the record of the fact to which they pertain. Once the copyright of records of facts is accepted, certain areas of copyright law become more certain. The areas of uncertainty in copyright law include:

¡ substantial copying; ¡ online access; and ¡ term of protection. If records of facts are protected by copyright, copying a single record of a fact without permission becomes an infringement of copyright. Copying records of facts is the main problem in downloading from online access. In addition, protection of access to the online database is required. This is already catered for in the UK by the Computer Misuse Act 1991. The term of protection of a data set becomes linked to the data items. Thus there could be a term of protection for each record of a fact which starts at that point in time when it is first recorded. The actual term of protection necessary and the terms of related data sets are discussed later in the chapter.

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

143

The transmission of data is another area of concern that is clarified by considering the record of a fact as property. A data message which is given to another party for transmission will not become the property of the transmitter unless the transaction included sale. Adaptation of records of facts, other than deriving data, does not affect the ownership of copyright of the work. Thus anything from a simple sorting of the work to its conversion to another format for transmission does not affect the underlying copyright. The transmitter of the message is therefore a courier, as would be the case for a physical package, whose service is paid for by the copyright owner. If the transmission is intercepted then a theft has occurred since the record of the fact is property. The copyright owner may seek redress in the courts against the thief or against the party transmitting the data for lack of security. The transmitter should be protected in law from unlawful access to its service, but this is not the function of copyright law. From what has been discussed on rights in records of facts, it can be seen that such records exist in digital and analogue form. In isolating and protecting the essence of data we are bound to protect data in whatever form they take. Thus the legal framework should apply to the general area of records of fact rather than to digital data alone. This has the added advantage that the legal framework should not be tied to the technology of a particular time. The obvious benefit of this approach is that it takes account of the movement of records of fact between its various forms and therefore takes away restrictions upon the user of data and clarifies the rights in derived products. From the description in this section we may summarise the requirements for rights in a data set as follows:

¡ copyright must extend to a record of facts; ¡ reproduction of a data set incidental to use must not be considered an infringement of copyright;

¡ derived works must be considered to be the result of the joint authorship of the data provider and the user;

¡ any joint author must be able to use his work without the requirement to gain the permission of the other joint authors;

¡ any joint author must have the right to realise his share of a derived work; ¡ any joint author must be able to have quick and easy settlement of a ‘royalty’ arrangement;

¡ the legal regime must clearly associate the copyright owner with his work; ¡ there must be no right of copyright owner or the creator of the work to object to derogatory treatment of a copyright work; and

¡ the act of transmission should not cause rights associated with access to data in copyright law.

Third party rights A number of rights in data may be referred to as third party rights. These third party rights relate to the following:

¡ ¡ ¡ ¡

data subjects’ rights; copying of works by private individuals; copying of works by educational establishments and libraries; and public access to government data.

144

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

Data subjects’ rights and copying by educational establishments are discussed further below.

Data subjects’ rights It is submitted (Larner, 1996) that, for a comprehensive legal framework for data, the data subject, or individual, could include legal entities in addition to people. Since it is proposed that records of facts may be owned, one might propose that the subject of the data set should have ownership rights in the data set. Jenkins (1989, p. 7) cites a US case in which the California court of appeal held that ‘an individual has property rights in unique cells and genetic information taken from his or her body’. One might also consider a parallel to this situation with geographic data. In the case of geographic data gathered by observation of the real world, the data subject is often the owner of the object not the object itself. Consequently, the data subject might be considered to be the owner of the fact but not the record of the fact. In the case of genetic codes noted above, the distinction between facts and records of facts has not been considered. Furthermore, it is questionable whether the ownership of facts relating to land and property should be possible. Few of the facts relating to a property are unique, for example, roof construction material. In addition, in so far as they may be observed from outside the property, it should be possible to record facts relating to a property except where such recording becomes harassment of the occupier. If individuals could stop the creation of records of fact, they would be restricting the free flow of ideas. However, the remuneration for the exploitation of such records is an issue separate from the ability to stop the records being made. A discussion similar to that of genetic codes, regarding rights in a work which records facets of a person, occurred in the determination of ownership of copyright in photographs. In the UK it was considered that the photographer owned the copyright in a photograph of another individual because the photographer arranged the composition. The model photographed was explicitly considered and excluded from ownership of the copyright in the photograph (Department of Trade and Industry, 1986, p. 7). Consequently, one can state that, in general, the property rights of records of fact that are created through observation should not involve similar rights for the data subject. However, a similar conclusion might not be justifiable for data provided by the individual in response to questions posed by a data collector. When the individual responds to questions, the record of the fact has originated from the data subject. Any use of the data for purposes other than those for which they were given invades the data subject’s ‘information privacy’. However, where the individual would be required by statute to provide those same data for another purpose one can argue that, with the exception of notification of the use, the individual has no grounds to object to that other use. Consequently one can argue that for statutory purposes unrestricted flow of data between users should be allowed. According to Hustinx (1993), the transparency of the processing that occurs and who holds which data are more important than consent by the data subject. In cases where the data provided are used for commercial purposes, one might argue that the subject has a right to remuneration. However, in the UK, the Lindop Committee (Home Department, 1978) held that the data subject’s remuneration usually comes within the provision of the service associated to the collection of data. For example, the discoveries about the genetic material described by Jenkins (1989, p. 7) related to an operation which the data subject required. Consequently the financial gam in exploiting the genetic data

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

145

reduces the financial burden of providing medical care. The same conclusion may be reached with respect to the provision of data to organisations supplying commercial services in general as long as the data supplied are not excessive with respect to the service provided. From the consideration of the case of Bernstein of Leigh (Baron) v. Skyways and General Ltd [1978], one might argue that the only limitation upon the subsequent use of data should occur when the use amounts to harassment. The judgement of what is harassment in the use of personal data, like the definition of privacy, varies. As noted by both the Younger and Lindop Committees, such judgements should be made according to the type of computer application and in a way that takes into account the rights of society in the efficient conduct of business. Where an individual is supplying data to an organisation whose sole purpose is to sell those data, then the individual is entitled to refuse to provide those data or to demand remuneration. The Younger Committee find one reason for giving the data subject ownership rights in personal data. The Committee (Home Department, 1972, p. 9) states that, ‘There is however one respect in which existing law seems evidently deficient, namely in the absence of any remedy for the use or dissemination of private information by someone who obtained it illegally or knew that it had been obtained illegally by another. Under the law of theft information as such cannot be stolen’. If records of facts were capable of being owned then personal data would be capable of being stolen. Consequently, if data were taken from an individual under false pretences there would be a right in law to pursue the offending party. However, this does not mean that the data subject should have the ownership right. They have sufficient redress by action against the data user for lack of security. In addition, the data user has the ability to take action for theft of the data by virtue of their ownership of facts and illicit access to the data by virtue of the Computer Misuse Act.

Education and libraries Lending by libraries, possibly even more than private copying, erodes copyright owners’ ability to profit from their work. Hugenholtz (1994, p. 37) states that, ‘As more and more works are available primarily or exclusively on-line, it is critical that researchers, students and other members of the general public have on-line opportunities equivalent to their current opportunities off-line to browse through copyrighted works in then- schools and public libraries’. However, online access to a library would be in direct competition to the sale by the copyright owner. Unlike the physical lending of a copy of a work, the number of ‘borrowers’ is not limited by the number of copies held by the library. If the public lending right is to extend to digital data then the number of copies available for lending must be limited. In the UK, as with all other fair dealing exemptions, there is a limitation on copying for the purpose of education. Geographic data are available from some suppliers at a discount. However, in some cases the terms of availability at educational prices do not reflect the required use. It is submitted that the terms of the use of data for educational purposes requires proactive adjudication by an independent body.

Term of protection Under the Copyright Act 1988, the duration of protection for a data set in the UK is uncertain. If the work is considered to be computer generated then the copyright lasts for

146

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

50 years. If the computer was only used as a tool in the creation of the work then the copyright lasts, with the exception of Crown Copyright, for the life of the author plus 50 years. It has been proposed that a new legal framework for data would have a different basis for duration of protection. The period of protection would run from the moment that a record of a fact was created. It has been proposed by some writers that the copyright of a data set should last for as long as that data set is in use (Metaxas, 1990). It has been held by the Supreme Court of the United States that a newspaper has a right under unfair competition law to control its news data whilst it is still ‘hot news’. This precedent, if applied to the commercial exploitation of geographic data, would result in the position adopted by Metaxas. The European Commission directive on the legal protection of databases does not adopt this position (see also Chapter 9). However, where sufficient change has occurred in the database the 15 year term of protection may be renewed. In the UK, the common law rights of the author were first temporally restricted to ensure the free flow of ideas. Given that the owning of records of facts would not restrict the free flow of ideas, particularly given a right to create derived products, it may be questioned whether the temporal restriction of the rights in a data set is necessary. Bainbridge (1994, p. 94) argues that moral rights should not be perpetual because it may not be clear in the future who holds the right. This could equally apply to the rights of the copyright owner. Consequently, if a perpetual right is accepted then it would be necessary to allow for certain data sets to be brought back into the market when their owner cannot be traced. Bringing data sets back into the market is in some respects analogous to curbing the tide of real property in the land market. If a temporal restriction is placed upon the protection of records of fact it should allow sufficient time to exploit the data set and thus encourage the creation of data sets in the first instance. Determining what constitutes an adequate period of protection for a data set is an economic question. Waldorf (1995) states that the half-life of a new map produced in the USA is one year. However, she admits that the half-life of the underlying data is a separate issue. In the USA the economic life of data is distorted by the availability of low cost, copyright-free government data. Waldorf quotes a cost of $300-$600 to produce a map in colour for inclusion in a travel guide. Similarly, road data for the USA may be leased from ETAK for $2,400 for the whole USA. Similar products in the UK would cost significantly more because the base data from which such products are derived must be costed into the production. It is clear that the UK approach to direct cost recovery by government departments is being pursued throughout the EU. An analysis of Ordnance Survey’s finances suggests that the digitisation of its largescale mapping cost in excess of £60 million. Such data sets are not subject to large amounts of change; consequently the period of protection is of paramount importance. Currently, the market for geographic data in digital form is only just developing. Ordnance Survey started its digitisation programme in the 1970s and completed it as recently as 1995. Her Majesty’s Land Registry is not due to complete the conversion of its records into digital form until after 1998. As noted by Hammond (1986, p. 345) there is a time lag between inventions and innovations in industrial products. Equally there is a time lag between the creation of new products and their uptake in the marketplace. Thus the period of protection in the early stages of the development of digital geographic data products is likely to be greater than that required for a mature information market. For Ordnance Survey, the period of protection would need to be 25 years just to cover the creation of its product. If some data sets become exposed to a wider audience then it is possible that they will be subject to different economic pressures. For example, the bundling of mapping software

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

147

with Microsoft Office might result in the wider distribution of certain mapping products. Such distribution is not yet a reality; consequently, if the period of protection for records of facts is not to be perpetual then it is essential that it at least allows for sufficient protection to the major investments necessary in large geographic data sets. An economic analysis of the data market is essential to determine the minimum period of protection necessary to encourage data creation.

Administration and adjudication of rights in records of facts Any legal framework for rights in records of fact must include the following:

¡ rights in data sets; ¡ methods of notifying rights in data sets and data items and the associated licence conditions; and

¡ methods of identifying owners and licensees. The first item has been dealt with by defining a two-level hierarchy of interests in data: the owner of the record of the fact and the licenser of the record of the fact. The system of rights proposed also simplifies licence conditions. However, the last two points in the list relate to the method of operating the system of rights and have the same need for simplicity and clarity as the rights themselves. It was seen in Table 11.2 that the desires for a legal framework for data included simplicity, clarity and minimal cost. The aspects of the system that were of concern included the administration and adjudication of rights in a data set. The proposed approach to rights in a data set simplifies the system of rights but still leaves the problems of their adjudication and administration. These problems comprise deciding when a data set infringes the rights in another work and clearly giving notice of rights in a data set to potential users. Copyright law works best when creativity leads to easily distinguishable works allowing copying to be easily detected. Works of geographic data must have a ‘sameness’ in order that they may be used. Thus the way in which copyright is enforced is not ideal for geographic data sets. A new legal framework should allow parallel exploitation of a source of facts whilst protecting unambiguously compilations of records of facts that are the same or similar. Patent protection is designed to work where expression of the idea leads to sameness. Patent works by grant of a monopoly to the first registered inventor. This can be a greater incentive to creativity. Another advantage of patent law is that the system of registration makes pursuing infringement simpler and cheaper. The difficulty of proving infringement and the cost of legal redress makes copyright protection difficult, with most cases being settled out of court. As a result, case precedent is difficult to find. The cost of taking to court a party infringing copyright is measured in tens of thousands of pounds for a simple case and hundreds of thousands for a complex one. In the UK, if a point of law requires clarification then the case will almost certainly go to the House of Lords; in the case of BLMC v. Armstrong Patents the total bill was in excess of £1 million. In 1986 the costs of the Copyright Tribunal were £600 per sitting day (Department of Trade and Industry, 1986, p. 71). Thus the a priori settlement of rights through registration could reduce costs. Whilst patent may simplify the law through registration, it does have disadvantages. There can only be one winner with the award of a monopoly and the benefit of parallel

148

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

exploitation of a ‘data source’ allowed by copyright is lost. The patent approach is to try to give absolute protection to the supplier, and this can stifle innovation through the fears of either inventing but getting beaten to registration or the research not creating an invention. In order to gain patent protection, a work should show ‘inventive step’. A work is considered to have inventive step if a panel of experts looking at an invention does not feel that it was an obvious extension of current knowledge. The inventive step required for patent registration is not appropriate for a data set since it must be in a standard form if it is to be useful. If a new legal framework to protect data is to be created then a hybrid of the patent and the copyright approach could be beneficial. This approach would achieve the goal stated above of parallel exploitation of a source of facts whilst protecting works that are the same or similar. It is argued that the approach taken by copyright law to adjudicating rights in works is inadequate. It is submitted that the legal framework should make a priori decisions about the rights in a data set. These rights would be documented in a system of registration of title to a data set. When registering a data set the title to the data set would be examined once and for all. This registration would then be maintained as rights are assigned. The system of registration and adjudication of title should be backed up by a system of adjudication of licence conditions. Application to a tribunal for a ruling should be possible before the need for litigation arises. Any licence or licensing scheme should be capable of submission for review of conditions or level of payments. Currently, the UK’s Copyright Tribunal may only consider licensing schemes of monopoly concerns. The Tribunal would also review a fair price for a derived product taking into account the value of the product from which the derived product was created. In turn the Tribunal would determine the proportion of the profit going to the supplier of the source data.

Compulsory licensing The compulsory licence is, in effect, the replacement of the exclusive rights of the copyright owner by a right to equitable remuneration. Currently, such compulsory licensing is a part of the blanket licensing schemes run by collecting societies or licensing schemes imposed by statute upon monopolies. According to Hoeren (1994, p. 60), blanket licensing is used by collecting societies to reduce the burden of data necessary regarding the interests in works or the number of users of a particular work. When records of facts are owned, a particular data set may be subject to a large number of rights. It is essential in these instances that the potential user of the data set may deal with it without having to spend large amounts of time and money ensuring that the data set is properly purchased. The purchaser would wish to have a simple contract that applies to the whole data set. The parties with ownership rights in the data set would wish any prospective purchaser to have notice of their interest in the data set. This might be achieved in a number of ways including metadata and registration. Metadata are already used to describe geographic data for a variety of purposes. In this case each part of the data set which is subject to the rights of a different owner would have metadata indicating this. Documenting a complex set of rights in a data set would be an additional burden on storage space that would bring no added value to the use for which the data set was purchased. The purchaser’s only interest is to ensure that all parties have received their part of the purchase price and royalties for additional copies made. A similar position arises when buying real property that is subject to a trust. The purchaser does not wish to know the details of the trust, only that the land is bought lawfully.

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

149

Real property registration of title indicates if a piece of land is subject to a trust. The register reveals sufficient detail to allow the purchaser to ensure that the purchase money is paid to two trustees, thus allowing the purchase to be performed with the minimum of effort but protecting the interests of the trust. It is submitted that registration of title would have a similar advantage for data sets. Potential purchasers may wish to acquire:

¡ ownership of the records of fact by assignment; or ¡ a licence to use the data set. Registration could ensure that notice was given of the owners’ interest to any party, whether that party wished to purchase the ownership rights in the data set or to license a copy of the data set for use. The registration of licensees of the data set would give notice of their rights to any potential assignee of ownership of the data set. The problem of complex interests in a work is not unique to data. Traditional copyright works in the performing arts are subject to a multiplicity of rights. It is this area of copyright where collective licensing is most prevalent. A number of different approaches have been taken in the EU to the administration of collective licensing. The UK and Italy represent the different extremes. Italy centralises all collective licensing in a single state agency, whilst the UK leaves the market to regulate itself. The registration of title to data would be a compromise between these two positions. Registration of title would relieve the user from the burden of finding the data set owners, allow economies of scale in giving notice of an interest and allow the operation of the free market in ownership of data sets. An open register would help to ensure that compulsory licences were complied with. In a number of cases, compulsory licensing might be necessary, including where relief has been sought from licensing conditions created by a monopoly. Furthermore, if the approach to charging for public sector data were to be changed, the registration system could easily allow that conditions were changed in the register in an open manner. This approach is a collective administration of rights rather than a collective licensing of rights.

Tribunal Vaver (1990, p. 116) quotes Michael Goldhaber as stating that ‘intellectual property laws…aid in the concentration of wealth and raise profits’. Goldhaber argues that this protection has the negative impact of forcing the reinvention of technology since ideas do not flow freely. He goes on to state that, “The innovation process should be as open and democratic as possible…. There should be quick response to negative impacts’. Vaver (1990, p. 110) states that if the author does not wish to disclose the work then copyright law and common law protection of confidential information turn into a ‘blunt weapon of censorship’. Vaver also states that ‘if the allocation of property rights is designed to encourage creativity and the free flow of ideas then when it does not achieve this the whole system of rights is unjustifiable’. The arguments put forward by Vaver and Goldhaber take an extreme view of the framework for protecting intellectual property. Erich Graham (1994, p. 109), of Bertelsman AG, put the opposing view when he stated that the ‘right to say “no” was the most important weapon for an author or publisher seeking to find the most attractive market’. However, allowing an author to say no in all conditions can lead to an abuse of monopoly position. It is when this abuse occurs that copyright stops the free flow of ideas. It is submitted that the property rights in records of fact and the rights in derived works proposed above avoids the main problems of monopoly abuse that exist under the current

150

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

copyright system. It is submitted that in allowing any joint owner of a work the right to deal in that work a precise use of compulsory licensing has been created which in effect stops single owners from abusing their position. However, as noted above, there is still a requirement for ease of access to binding arbitration regarding the shares in a jointly authored work. It is submitted that the arbitration of licences and licensing schemes should be carried out by a tribunal.

Use restrictions on ownership rights Hoeren (1994, p. 61) states that a legal system for digital products ‘has to be drafted to the effect that the licensee can report the purported use of the work in order to get rid of the moral rights problem’. The proposed framework of rights would not require the notification of use for moral rights purposes since the objection to derogatory treatment would not be a part of the system of rights. However, the relationship of purchase to use of the data set does touch upon the suggested amendments to the processing of personal data. The proposed solution to the problems of abuse of data has been to ban those uses that should be unlawful in any circumstances and to control those uses which may be abusive if performed in certain ways. Consequently, the control of applications could interact with registration of title in two ways:

¡ the assessment of the data set on first registration and registration of use inhibitions; and

¡ the assessment of the applications within which a licensee will use a data set for potential abuses. The latter procedure is already carried out in the UK by the Office of the Data Protection Registrar. Thus, for the UK at least, there is no increase in administrative overheads. Through guidelines and case precedent, regulations are being built up for certain uses of data. It has been argued by Larner (1996) that the UK’s system of data protection should register the data set and use inhibitions on application types rather than specific uses of data sets. Thus, if the title register included all personal data, in addition to the creation of that title register, the Data Protection Register would be reduced to a register of applications. Furthermore, the two tribunal’s currently used in the UK could be reduced to one. This tribunal would take on a more positive and proactive role in giving a user a clear licence to use a data set without having to declare each and every purpose. The proposed approach would release the Data Protection Registrar from the burden of maintaining a register and release the time of data protection officials to controlling specific computer applications. More importantly, it would separate the role of policing the use of applications from adjudication among owners, users and subjects. Given the current approach to registering personal data, the burden of registration in the personal data protection system would involve less work and have the attraction of protecting the rights of the party who registers. From the different circumstances of registration or consultation of the register described above, one can say that the register will have to record the following:

¡ ¡ ¡ ¡

data set; owners of the data set; licensed users of the data set; and inhibitions on use of the data set.

LEGAL AND INSTITUTIONAL ISSUES TO BE RESOLVED

151

Conclusion It has been submitted that for the integration of European data it is necessary to define a clear system of rights in data. Before such a system of rights may be created it is necessary to have a universally accepted definition of data for the purposes of legal analysis. Furthermore, it is necessary to define what a system of rights should do separately from an analysis of what the current law does. The agreement upon what a system of rights should achieve must involve representatives of all types of party to data handling—owners, users and subjects. In turn, the system in which rights will be enforced must also be created. Once again the interests of all parties must be assessed with respect to enforcement of rights. This chapter has discussed a small part of a proposed system of rights. That system is closely linked to the UK experience. Further discussion is required for any system of rights to be agreed at the UK level, let alone for a harmonised European system. Once the desired qualities of a system of rights have been agreed then current European law may be analysed for its effectiveness. It has not been possible within the confines of this chapter to review European law with respect to the proposed system of rights. The proposed system of rights has been discussed in the language of ‘copyright’. This does not mean that a new set of rights must be framed in copyright law. Alternative methods of enforcing rights in data have been proposed. Technical devices, built into either hardware or software, may also be used to control rights in data. These devices may be seen either as an alternative or as a supplement to the proposed system of adjudication and administration. Hoeren (1994, p. 61) states that ‘technical devices have no disadvantages per se; they are instead mandatory for solving multimedia licensing problems’. Lewis (1987, p. 8) echoes Hoeren’s sentiments when he states that ‘lawyers will never be able to provide a practical answer to the prevention of breach of copyright in this area and that we must look to technical solutions if we wish to safeguard copyrighted material’. However, such technical devices are open to abuse. If rights in data are not defined with respect to all types of party then technical devices will merely enforce the suppliers’ view of rights. Given universal access through a computer network, technical devices created for the benefit of the data owner according to their interpretation of what is a legitimate use would become a stronger device of censorship than that used by Henry VIII in the sixteenth century. The proposed system of enforcement of rights has been based upon a system of registration. However, such a proposal requires further research. The economics of the data market in general and the geographic data market in particular will determine the viability of any system of registration.

REFERENCES AGI (1992). Report by the Copyright Committee. London: Association for Geographic Information. AGI (1993). Ordnance Survey Charging Round table, Final Report and Recommendations. London: Association for Geographic Information. BAINBRIDGE, D.I. (1994). Intellectual Property (2nd edition). London: Pitman. DEPARTMENT OF TRADE AND INDUSTRY (1986). Intellectual Property and Innovation, Cmnd 9712. London: HMSO. GRAHAM, E. (1994). Minutes of the meeting of the Legal Advisory Board of the Commission of the European Communities, Proceedings of the Conference on Legal Aspects of Multimedia and GIS, 27– 28 October 1994. Luxembourg: Commission of the European Communities.

152

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

HAMMOND, R.G. (1986). The misappropriation of commercial information in the computer age, Canadian Bar Review, 64, 342–73. HOEREN, T. (1994). Legal aspects of multimedia—long term solutions, Proceedings of the Conference on Legal Aspects of Multimedia and GIS, 27–28 October 1994, pp. 52–62. Luxembourg: Commission of the European Communities. HOME DEPARTMENT (1972). Report of the Committee on Privacy, Cmnd 5012. London: HMSO. HOME DEPARTMENT (1978). Report of the Committee on Data Protection, Cmnd 7341. London: HMSO. HUGENHOLTZ, P.B. (1994). Copyright problems of electronic document delivery, Proceedings of the Conference on Legal Aspects of Multimedia and GIS, 27–28 October 1994, pp. 34–42. Luxembourg: Commission of the European Communities. HUSTINX, P. (1993). Privacy and data protection—balances and perspectives, Conference of the Parliamentary Information Technology Committee: The Emerging Information Society—The Public, Private and Personal Policy Issues, 20 October 1993. London: PITCOM. (mimeo) JENKINS, L.D. (1989). The legal dimensions of information—patents, Proceedings of the Institute of Information Scientists Annual Conference: Where the Book Stops, pp. 3–8. London: ASLIB. LARNER, A.G. (1992). Digital maps: what you see is not what you get, CS Monthly—Land and Hydrographie Survey, 1(6), 4–7. LARNER, A.G. (1996). The legal and institutional restrictions on the handling of digital land related data in the UK. PhD thesis, University of East London. LEGAL ADVISORY BOARD (1994). Proceedings of the Conference on Legal Aspects of Multimedia and GIS, 27–28 October 1994, Luxembourg: Commission of the European Communities. LEWIS, D.P. (1987). Copyright aspects of databases, Computer Law and Practice, September/ October, 2–8. METAXAS, G. (1990). Protection of databases: quietly steering in the wrong direction, European Intellectual Property Review, 7, 227–34. NIMMER, R.T. and KRAUTHAUS, P.A. (1993). Information as property databases and commercial property, International Journal of Law and Information Technology, 1(1), 3–34. VAVER, D. (1990). Intellectual property today: of myths and paradoxes, Canadian Bar Review, 69, 98–125. WALDORF, S.P. (1995). Cartography for consumers: meeting new demands with digital technology, addendum to Proceedings of the International Cartographic Conference. Barcelona: ICA.

Statutes and legal cases Bernstein of Leigh (Baron) v. Skyways and General Ltd [1978] 1 QB 479. Kenrick v. Lawrence [1890] 25 QBD 99.

CHAPTER TWELVE

Intellectual property rights in disseminating digital geographic data, products and services: conflicts and commonalties among EU and US approaches HARLAN ONSRUD AND XAVIER LOPEZ

Introduction It is now generally accepted by academics, government and the business community that information has become one of our most important resources for the generation of wealth and power. The battle over control of information systems and the content carried by them is in full fray yet the grab for intellectual property rights has been largely ignored by the press and, as a result, seemingly has gone unnoticed by the general public. While the.press focuses on new offerings on the Internet, the latest computer products, and issues such as cyberporn, corporations around the world have been lobbying their governments for expansion and extension of intellectual property laws. As noted by a US law professor in a recent editorial in the New York Times, ‘Governments are complying, granting monopolies over information and information products that make the monopolies of the 19th-century robber barons look like penny-ante operations’ (Boyle, 1996). Particularly bothersome in the spatial data context is the ‘sell out’ by government agencies to corporate ‘partners’ of intellectual property rights in core data sets upon which value added products and services might otherwise be built by a wide diversity of private sector innovators and distributed through a diversity of channels. Spatial data sets are expensive to create and are now very easy to copy. We need intellectual property rights to protect them. Without some form of protection there would be little incentive in the private sector to create them and make them available to the public. Proponents for expanded protection argue that the more we protect intellectual property rights in data sets, software and information products, the greater the incentive there is for innovators to create them and offer them up for use by all of us in society. However, there are limits to this theory in a practical world. Innovators need an adequate amount of raw material from which to create and innovate. Data and information themselves are this raw material. The market cannot operate without an adequate flow of information. 153

154

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

In granting data monopolies or establishing government practices that encourage the creation of private sector oligopolies, we reduce the public domain such that creators no longer have the raw materials to create. By imposing monopolistic conditions on parties that would like to add value to raw data and sell the resulting product, or by imposing monopolistic pricing, the raw material becomes so burdened with the rights of others that the incentive for innovativeness is lost and economies are stifled. Thus, society should protect intellectual property rights in privately created data sets but only within an overall context of enlarging the public domain of raw information materials such that innovations are encouraged and within a context of providing deference to new and expanding innovations over old innovations. The appropriate role of the academic research community is to continually question the logic and validity of arguments presented by parties to important societal disputes. In evidencing the truth or falseness of claims, academics often strive to collect evidence on the ramifications, for instance, of following one information policy approach over another. Their work is typically reviewed by peers with an expectation of full disclosure of study and survey methods. Considered reflection and analysis is the expected norm. Further, when academics identify interests not represented in a social policy debate, they often take on the role of exposing and articulating the interests of those parties that may be disfranchised. Thus, our bias in the following is to take a citizen advocacy role. We feel that many of the long-term interests of consumers and general members of society have not yet been heard clearly in the current debates over information policy choices. We recognise that others as well come to the table with their own respective biases. Some of the arguments presented and positions taken in this article admittedly may not be in the short-term economic best interests of some government agencies and private corporations. Academic researchers often are critical of ‘innovative progressive practices’ by government agencies and industry and in turn academics are criticised as ‘lacking understanding of the realities of changing political and economic conditions’. Academic researchers counterargue that government agencies and private industry have more than ample resources to adequately represent and promote their own positions and interests. In addition, in addressing the issue of protecting intellectual property rights in geographic data sets we believe there are substantial opportunities for reasonable compromises that support the primary policy goals of most stakeholders, including the general public, private businesses, and government agencies. For instance, we believe that rational models now exist in the USA for instituting geographic data dissemination policies in US local governments that accommodate both open access to geographic data for local businesses and citizens and yet also meet the primary revenue generation objectives of local governments (Onsrud et al., 1996). We believe similar compromises supporting most of the primary policy goals of most stakeholders may also be achieved relative to sharing and commercial activity in digital geographic data at an international level.

Intellectual property rights in geographic information The use of copyright law to control the dissemination of government information dates back to British copyright law which originated in the Statute of Anne, enacted in 1709 (Statute of Anne, 1978). Over the past three centuries, copyright has evolved from a state-enforced form of censorship into a constitutionally protected doctrine to ‘promote the progress of science and the useful arts’ (Goldstein, 1994, p. 19). Over the same period, copyright law has had to adapt to the various forms of fixation used in the

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

155

communication of expression, from printing press, to photography, film, video, radio and now databases. As explained in earlier chapters, one objective of copyright law is to encourage expression of ideas in tangible form so that the ideas become accessible to and can benefit the community at large. Copyright compensates creators of original expression as an incentive for them to continue to bring forth knowledge and information that others in the community might be able to exploit for social or commercial gain. The intent of copyright is to protect expression but not to protect ideas. The ideas in a work protected by copyright can and should be used without compensation to the creator. However, those expressions imbued with originality by the author are protected. Differences among the copyright laws of various nations have resulted from a wide range of interpretations that nations have developed for the concept of originality.

Government copyright Copyright of government information is common throughout North American and Europe. The USA is one of the few industrialised countries which expressly forbids federal agencies from imposing copyright against its citizens (with a few exceptions), thereby placing these information resources in the public domain. Most state and local governments in the USA have the option of imposing copyright in their public records if they choose to do so. In other countries, copyright may be applied to all works of government. However, even in nations allowing broad copyright in national government data sets, there are obvious variations in the implementation and enforcement of copyright. For instance, the UK favours strong protection while France and Canada attempt to balance between the needs of protecting government works and the needs of citizens to access public information. Thus, in the latter countries, works of a regulatory or policy-making character are often excluded from government copyright. The conversion of public information to digital form makes government data more valuable and therefore raises the stakes in the information policy debate. The conversion from print to data sets has led many government departments around the world, including national mapping and statistical agencies, to assert their copyright, apparently to secure the potential revenues from data sales. Potential revenues from the sale of spatial data sets may be attractive incentives to government copyright holders, just as they would be to other holders of a valuable information asset. Owing to their dominant positions and fiscal incentives to do so, it is very likely that most government agencies will choose in their own best interests rather than in the interests of citizens generally in deciding whether to impose copyright and to what extent on the geographic data sets that the agency has created. Thus such issues should be addressed by public policy-makers and elected officials. The challenge for public policy-makers is to determine the extent to which copyright or other intellectual property controls should be utilised by government agencies in order to generate revenues and to determine the economic and social costs to the public interest and commercial innovation for each of the policy routes that might be chosen.

Copyright of electronic data sets in the USA after Feist Electronic data sets, also referred to as compilations, are protected under the US Copyright Act 1976 (USCA). The Act defines compilations as the ‘collection and assembling of pre-

156

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

existing materials or data that are selected, coordinated, or arranged in such a way that the resulting work constitutes an original work of authorship’ (17 USCA, section 101). The level of originality required in order for a work to be protected under the Act is very low in the USA. However, protection extends only to the originality aspects of the work. Thus for geographic data sets, creativity aspects of their selection or arrangement are entitled to copyright protection whereas their factual content is not. US copyright law represents a careful attempt at balancing the rights of creators with those of the users. Its central purpose is to stimulate learning, research, education and the creation of new works of art and science. As Justice O’Conner wrote for the US Supreme Court: the primary objective of copyright is not to reward authors, but to promote science and useful arts. To this end, copyright assures authors the right to their original expression, but encourages others to build freely upon the ideas and information conveyed by a work…. This result is neither unfair nor unfortunate. It is the means by which copyright advances the progress of science and art. (Feist Publications v. Rural Telephone Service Co., 1991) Since the production of information resources is the fundamental goal of copyright laws it is necessary to understand better how much data government agencies and publication offices are willing to make available in electronic format. If intellectual property is too lax, there may be inadequate incentives to produce additional information works. However, if protection is too rigid, it may impede the free flow and fair use of information (Varian, 1995). In addition, the rights of copyright owners must be broad enough to provide a fair return on their work. The 1991 Supreme Court ruling, Feist Publications v. Rural Telephone Service Co., held that white-page telephone directories were not copyrightable since these compilations were insufficiently original to warrant protection. This is relevant to spatial databases, since a GIS data set is more akin to a directory than it is to a ‘pictorial, graphic or sculptural work’ (17 USCA, section 101, West suppl., 1992). Moreover, GIS data sets are unique in that a database management system (DBMS) allows the user to separate easily the factual content from the arrangement without infringing copyright. This was previously far more difficult with hardcopy maps. The Feist decision demonstrated that white-page telephone directories were not sufficiently original to warrant protection. The underlying meaning of the Feist ruling is that a compilation, such as a directory or a factual database, to be protected by copyright, must be a product of a minimally creative selection, coordination or arrangement process, rather than mere effort. A decision from the Southern District of Texas regarding copyright protection for parcel maps concluded that ‘maps express the only pictorial presentation that could result from a correct interpretation of the legal descriptions and other factual information relied upon by the plaintiffs in producing the maps’ (Mason v. Montgomery Data, Inc., 1991). The consequence of this district court decision was that creative expression and selection anywhere on the maps was protected. However, the factual information itself was not protected. If one considers a statistical table or data set of coordinates and features, copyright protects the arrangement and labelling of columns and variables, but not the numbers or statistical values represented. Nevertheless, the courts are likely to be sensitive to the livelihood of data set creators whose competitors pirate a copyrighted data set and sell it for profit while providing little or no enhancements. Hence, the use of misappropriation arguments will be relevant in cases of blatant data set piracy. However, if a competitor gains access to the raw factual

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

157

data and significantly enhances these data by combining them with other data or improving their quality, the original data set creator may be hard pressed to win a copyright infringement claim (Raskind, 1991, p. 347). The importance of the Feist decision is that it has confirmed a US policy of stimulating competition and innovation in information markets. Feist appears to introduce a competitive thrust into copyright law that will require data set creators to continually enhance their product in order to stay one step ahead of their competition.

Copyright of geographic data Even though a spatial database might cost hundreds of thousands of dollars to compile, it is typically far more difficult to use the law to support copyright in a GIS database than it is in, for instance, a novel. One reason for this is that spatial data are largely factual in nature and facts are not subject to copyright under the Berne Convention (see Chapter 9). Facts, algorithms, physical truths and ideas exist for everyone’s use. It is difficult to argue that the outline of a building, the bounds of a land parcel or a contour line on a map are expressions of originality. Any other person or sensor attempting to represent these physical facts would have little choice but to do so in much the same way (Kern River Gas Transmission Corp. v. Coastal Corporation, 1990). To represent the features by other than points, lines, polygons or image bits would make the representation non-standard, greatly decrease the value to others and make the data useless or cumbersome for computer processing (Onsrud and Reis, 1995). The ‘merger doctrine’ is another well established legal principle under US copyright law. Under the doctrine, copyright protection will not be granted if there is only one or a small number of ways to express an idea. Thus, by example, even though a building outline data set might reach the required ‘modicum of originality’ in order to fall within the realm of copyright protection, copyright protection would be denied because the expression is one with the facts. The fact or idea of how far a building extends and the expression of that fact or idea are merged. The expression of the building limit as a line is largely inseparable from the fact or idea of the limit because expressing the limit in any other form than a line would be highly impractical in a real world. The argument might be made that a building line expressed on a map is an author’s opinion of where the limits of a building are, rather than an expression of factual location subject to some degree of error. However, if many mappers independently collecting the same building limits arrive at the same expression for the limits of the building within a reasonable degree of certainty for most potential uses of the information, one may only logically surmise in a practical world that the information provided is more in the nature of factual information than opinion. Thus, clearly, a solid line representation of a building limit fully meets the merger doctrine test. Similarly, a contour line above a standard and widely used datum collected and expressed at a standard and widely used mapping scale exhibiting a reasonable degree of accuracy for many practical purposes would also appear to meet the merger doctrine test. The intent of the merger doctrine test is to help to ensure that ideas and standard factual data are kept within the public domain for all others in society to productively use and build upon. Although the doctrine is well established under US law, it is not as evident under the copyright laws of European nations (Samuelson, 1996). In the USA, if one is allowed to gather data through indirect copying such as recompiling the factual listings in a telephone directory from original data or taking aerial photographs

158

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

of an area in order to create outlines of buildings already shown on someone else’s map, then one is not prevented by copyright from copying the facts from an existing source (Feist Publications v. Rural Telephone Service Co., 1991). However, there is still the potential that other laws such as misappropriation or unfair competition laws might be imposed against a person that takes the data from another individual’s compilation rather than gathering the facts independently. Even if a copyright claimant argues the existence of originality in the selection, coordination or arrangement of a data compilation, the typical GIS user is primarily interested in the data themselves and not in the originality aspects of the data compilation (Lopez, 1993). They wish to use segments of the factual data to do their own original work. Thus if one extracts only factual information from a geographic data set to create a new arrangement, it is difficult to argue that the individual has copied another’s original expression. Under US law, at least, Feist suggests that completeness or comprehensiveness in the collection of a set of facts (i.e. all telephone listings in a community) makes the collection quite ordinary rather than ‘original’ and being first to compile a set of facts fails by itself to establish a copyright (i.e. rejection of the ‘sweat of the brow’ copyright theory). Data set ownership interests are further complicated in GIS environments because many of the data currently being compiled in them are being copied from existing paper maps; some of which are obviously in the public domain but many others of which involve potential pre-existing copyright interests. The test used in the USA to determine whether newly developed software has violated the copyright of existing software is set forth in Computer Associates v. Altai (Samuelson, 1996). If this test were similarly applied to spatial data sets it would first require the identification of unprotectable elements of a data set. These elements would include factual information but would also include data expressions dictated by efficiency considerations or by external considerations such as making the data set compatible with other data sets or software. These elements are not protected by copyright and are filtered out. This step of the Altai test suggests that the more standardised a data set is the less subject to copyright it should be. What remains after the filtering process are the ‘golden nuggets of expression’. It is these golden nuggets that would then be compared to the potentially infringing data set to determine whether those aspects of expression are present and, if so, whether they were illicitly copied. Copyright laws of the European nations tend to be painted with a broader brush stroke where protecting software is concerned. If found to be protected by copyright, that protection typically extends to greater than the ‘golden nuggets of expression’. At the current time there is no indication that the Altai test will be formally applied by the courts to data sets in the USA, but the principles supported by the test seem to comport with current copyright law as applied to data sets there. Because the Berne Convention and other international copyright conventions extend only to original works and not to facts, several legislative proposals have been proposed within the EU in regard to protecting the neighbouring rights of creators of works. The basis of the underlying theory is similar to the familiar ‘sweat of the brow’ theory recently rejected by the Supreme Court in the USA. Other theories being considered include moral rights, unfair competition, and sui generis protection (CEC, 1995a).

Contract law as applied to GIS data sets Given the limitations of copyright law for protecting spatial data sets, organisations that intend to maintain a proprietary interest in their information resources are increasingly

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

159

turning to contract law and the use of signed licence agreements to control the use and duplication of data sets. A licence is a legally binding contract whereby the terms between a database supplier and user can be agreed upon and enforced through the principles of contract law. Licences define the rights and privileges of the parties more specifically than would the copyright statute alone. These agreements generally determine limitations on duplication, resale and derivative products. Penalties for violation of the contract terms are defined before they occur, thus reducing the possibility of litigation to achieve enforcement. The restrictions elaborated in the contract typically limit uses, users, sharing arrangements and downstream dissemination of the data. Furthermore, licensing copyright works allows authors to receive economic gain at privately negotiated prices (Goldstein, 1977). The wording in contracts or licences is often also used to decrease liability exposure. Such language typically spells out the data supplier’s efforts to notify the user of the limitations of the database and gives notice of the remedies available to dissatisfied customers or those that are damaged through dependence on the data. Licensing agreements are quickly becoming the preferred means of achieving control over the use and reproduction of spatial data sets by suppliers. Contract law protects data set suppliers from misuse and copying of the arrangement and the factual content of a data set. These agreements provide data suppliers with a mechanism to protect the contents of factual data sets—an act not permitted under most copyright laws. A drawback of licences from the user’s perspective is that many contractual licences now in use for the acquisition of works in digital form fail to take into account legitimate uses under the ‘fair use’ exception to the copyright law (Litman, 1995). A drawback of contracts and licences from the data vendor’s perspective is that they may provide little protection against third parties that gain access to the data but have no contractual relation with the supplier. Unfair appropriation laws may be the only resort under such circumstances and may have little applicability in many instances. The motivation for data vendors to combine copyright with contractual licences is understandable: database development is expensive, copying is easy, and data suppliers wish to protect their investment and assure revenues. However, these restrictions can have serious implications on the general user community if they are imposed by government GIS agencies. At issue is the fact that suppliers, particularly in the public sector, remove the access to government information principle from the realm of public policy and reinterpret it from the perspective of business interests and principles. If not applied fairly, licences imposed by public agencies can potentially restrict the resale or value added activities of commercial and non-profit organisations. If users are forced to sign away rights to public data which were otherwise available under copyright law, the imposition of such language may raise ‘lack of consideration’ or ‘lack of just compensation’ issues under contract law principles. If found lacking, the contract may be invalid. Furthermore, licences and contracts tend to reinforce monopolistic tendencies of government information suppliers. This results from a captured market and the elimination of alternative information sources owing to the up-front sunk costs at public expense involved in government mapping activities.

US federal government copyright law setting US federal policies have been developed to address the public need for information (McClure et al., 1989; Hernon and McClure, 1993). US public information principles are

160

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

based upon an attempt to guarantee broad access to information as a precondition to economic and political opportunity. US policies have been derived from four broad motives:

¡ ¡ ¡ ¡

to encourage public education and enlightenment; to protect intellectual property rights; to assist economic development; and to protect national security. (Ballard et al., 1989, p. 86)

The intellectual property clause of the Constitution furthers these goals by requiring federally funded science to contribute to the public good (McClure et al., 1989, p. 152). US principles of open government are further enhanced by section 105 of the US Copyright Act which precludes copyright protection for works of the Federal government. The fundamental reason for not allowing federal agencies to copyright public information resources was the fundamental belief that government copyright is the antithesis of open access whereby an informed citizenry can check official abuses. However, other economic values are at work, primarily that individuals ought to be able to derive benefit from public goods (such as public information), and that education (increased access to information) is inherently a good in its own right (US Congress, 1986).

Freedom of Information Act (FOIA) The US Freedom of Information Act (FOIA) was passed in 1966, and strengthened in 1974, followed in 1976 by the Sunshine Act. Until the enactment of FOIA there was no statutory mandate compelling government to release information to the public (Branscomb, 1994, p. 167). The US federal government recognised the importance of informing its citizens by eliminating constraints to its access (price, intellectual property restrictions, etc.) through the enactment of FOIA (5 USCA, section 552, 1986). The Supreme Court has observed that ‘The basic purpose of FOIA is to ensure an informed citizenry, vital to the functioning of a democratic society, needed to check against corruption and to hold the governors accountable to the governed’ (NLRB v. Robblns Tire & Rubber Co., 1978). Hence, FOIA is viewed as an adjunct to the constitutional elements of a democratic system (Cooper, 1986, p. 622). Although FOIA does not specifically identify data sets as a government record, the federal courts have consistently held that computer records are public records for the purposes of FOIA (Yeager v. Drug Enforcement Administration, 1982).

Copyright and related administrative law as applied to federal agencies Government copyright The laws of the USA provide that copyright protection is not available for any works of the federal government (17 USCA, section 104, 1988). Thus, US government public agency records are regarded as being in the public domain. The federal government does reserve the right to enforce copyright of its public records against foreign users if it chooses to do so. Within the USA copyright may be imposed by state and local governments. Paperwork Reduction Act The US Office of Management and Budget is authorised to set federal information policy through the Paperwork Reduction Act and Circular A-130. The Paperwork Reduction Act

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

161

1995 (PRA) is the most comprehensive of the statutes pertaining to the dissemination of federal information resources (Perritt, 1995). The Act expressly mentions electronic information collection and dissemination techniques and obligates the director of the Office of Management and Budget (OMB) to develop government-wide policies for co-ordinating data acquisition requests, data use and information dissemination policies. It has been designed to complement the Freedom of Information Act by having government actively disseminate its information resources rather than rely on cumbersome administrative FOIA. The use of an electronic Government Information Locator Service (GILS) as a central computer interface and retrieval system is central to its dissemination objectives (Office of the President, 1994a). The purpose of GILS is to use the nation’s emerging national information infrastructure to help the public and agencies to locate and access information throughout the US government. OMB Circular A-130 The Office of Management and Budget directs federal government information dissemination policy through OMB Circular A-130 (OMB, 1992). This directive establishes the administrative rules and guidelines governing the dissemination of all federal government information (17 USCA, section 102). The most recent update of the Circular explicitly recommends that federal information resources be disseminated at the marginal cost of dissemination in order to encourage access and use through a diversity of channels. OMB policy clarifies the unique character of government information as a public good, which is statutorily exempted from copyright protection and is required to be disclosed upon request under the FOIA. The 1992 version of OMB Circular A-130 stipulates: In order to minimise the total cost and maximise the usefulness of government information, the expected public and private benefits derived from government information should exceed the public and private costs of the information, recognising that the benefits to be derived from government information may not always be quantifiable…. [U]ser charges higher than the cost of dissemination may be a barrier to public access. Given that the government has already incurred the costs of creating and processing the information for government purposes, the economic benefit to society is maximised when government information is publicly disseminated at the cost of dissemination…. However, where agencies provide custom tailored information services to specific individuals or groups, full cost-recovery is appropriate. (OMB, 1992) Circular A-130 and the Paperwork Reduction Act share two fundamental objectives:

¡ to improve federal agency electronic information management; and ¡ to ensure that public access provisions are built into the development of federal projects and programs. The US Copyright Act, FOIA, Circular A-130 and PRA have evolved over the past two decades to complement each other in order to ensure that government records, information and databases are more readily available to the public and private sectors in a manner which meets national priorities. National Spatial Data Infrastructure initiative Recognising the importance of federal action in co-ordinating the complex milieu of federal information handling activities, President Clinton signed an executive order in 1994

162

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

directing all federal agencies to contribute to the development of the National Spatial Data Infrastructure (NSDI) (Office of the President, 1994b). This presidential executive order lays out key activities that federal agencies must conduct in conjunction with state and local governments, academia and the private sector to ensure the evolution and growth of the NSDI. These key activities illustrate that both the government and the private sector must become major players, and in some cases partners, if government is to take economic advantage of its technical information resources. Indeed, in the areas of its primary US technological leadership—defence, space, agriculture, software and environmental technologies—the record is clearly one of an integrated information management capacity in which the roles and responsibilities of all sectors are closely intertwined. Furthermore, in 1990, a US Congressional report noted that in order to maintain a leadership position, effective government information policies will be needed to enhance a nation’s technical innovation and global competitiveness (US Congress, 1990). Likewise, the president of the National Academy of Public Administration recently noted that, [information] is pivotal to the vitality and productivity of government services and the nation’s economic competitiveness. At issue is whether we can use information technology effectively to empower government, the private sector, and citizens alike. The complexity of today’s world demands that the public and private sectors not only learn to master this tool, but also work cooperatively to maximise the national benefits. (NAPA, 1993) It appears that US policy-makers are now beginning to recognise the importance and economic benefits of promoting a robust spatial information infrastructure through the application of information policy mechanisms, as opposed to direct subsidies. However, the question remains: what is the appropriate information policy model that will propel the creation of a national spatial data infrastructure into the next millennium?

European Union copyright setting Freedom of access to environmental information With the formation of the EU, member states are increasingly being bound by legislation and decision-making made by the European Union (EU). The EU adopted in June 1990 the Directive on the Freedom of Access to Information on the Environment (Hallo and Roderick, 1995). This directive guarantees an individual the right to information held by public authorities about the environment. Information includes existing data collected or prepared by such authorities and which are contained in written documents, data banks or visual recordings. Although this directive is intended to make governments more open in their provision of environmental registers, the requirement to provide maps, digital map products or earth observation data is uncertain.

EU copyright The differences and anomalies of copyright law that exist in the EU member states have prompted the EU to establish a minimum set of standards for copyright law which would harmonise protection throughout the Union. As a result, the EU has recently adopted a directive for the legal protection of databases (CEC, 1996a).

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

163

The Commission feels that a consistent, secure and stable legal regime is necessary in the EU so that database creators and operators can compete on equal terms with their leading rivals in the world information market. The nationally disparate terms of copyright protection can be a barrier to the free movement of goods—a cornerstone of the EU. The EU’s two-tier approach provides full copyright protection to databases meeting the necessary originality criteria related to the selection or arrangement of a database; and an additional neighbouring right on the factual contents of the database. The directive provides protection for the database arrangement for 70 years. In addition, the contents of factual compilations (i.e. those ineligible for copyright owing to the lack of creative expression) would be protected for 10 years against unfair copying. The stated intent of the directive is to serve as a possible basis for international accommodation of new information products currently ineligible for traditional copyright protection. While most EC directives have had a primary goal of reconciling existing laws across nations, the Directive on the Legal Protection of Databases broadens the protection of factual compilations which are not eligible for protection under traditional copyright or other existing laws. The EU’s proposal on database protection intends to provide legal protection for these factual or low-originality compilations for up to 10 years. A forerunner of the directive addressed impediments to accessing databases arising from monopoly suppliers. It recommended that voluntary licence agreements be used to ensure access to proprietary information on fair and non-discriminatory terms. This compulsory licensing requirement would have required compulsory licences on all monopoly suppliers of government information in order to encourage value added commercial exploitation and competition. However, the compulsory licensing requirement was dropped from the final version. As a result, the final directive appears to overprotect database owners without balancing the need to provide compulsory licences or provide fair use rights to users that wish to acquire and use data from exclusive suppliers.

Administrative guidelines In light of the growing use of copyright by government agencies, the CEC Guidelines for Improving the Synergy between the Public and Private Sectors in the Information Market stresses that EU member states’ efforts to copyright their information resources may conflict with stated EU principles associated with a right of access to government information (CEC, 1993). The guidelines are the outcome of an EU project that is currently examining how member states assert principles associated with a general right of access to government information. It is being undertaken by the EU’s DGXIII Legal Advisory Board (LAB). Preliminary findings from this project suggest that greater openness to government information is necessary in order to encourage the synergy between the public and private sectors needed to establish an EU information market and to reduce possible distortions of competition in the European market for public sector information (Burkert, 1992). The EU guidelines emphasise that such openness is necessary to encourage synergy between the public and private sectors in establishing an EU information market and to reduce possible distortions of competition in the information market for public sector information. The CEC Synergy Study encourages governments to use their discretion to minimise public sector copyright. This report goes on to stress that the public sector should not indulge in discriminatory pricing or dissemination practices that would affect fair competition and be contrary to EU competition law. Given the importance of

164

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

government copyright on the development of a global spatial data infrastructure, better understanding of the impacts of these forms of control on public access and industry development is necessary. Since the EU Synergy guidelines have no sanctions attached, they have not been given attention by the member states (Burkert, 1992; Beasley, 1995). However, the important role of government information in the EU, coupled with the emerging European information infrastructure policy efforts, discussed below, are likely to place national information policies under greater scrutiny (CEC, 1995b; CEC, 1996b). The EU has embarked on a number of strategic initiatives promoting the development of an EU-wide geographic information markets. The EU efforts are driven less by ideology and more by the pragmatic goal of exploiting government information resources to stimulate the public and private information sectors. A principal goal of this activity is to stimulate innovation in the commercial information sector by making government information more readily available and by reducing the impediments to access (Policy Studies Institute, 1995). Although a political component is inherent in the EU proposals, the objectives are closely aligned to industrial policy and strengthening the competitiveness of European information industries. The EU initiatives are attempting to develop policies of access which exploit public sector information, link sources of European public sector information and make better use of content resources in the public sector. Specific programmes include:

¡ 1NFO2000: An EU programme to stimulate the development of a European multi-

¡

media content industry and to encourage the use of multimedia content in the emerging information society. One of its principal action lines is to better exploit Europe’s public sector information in order to stimulate interconnection throughout Europe and to bolster the competitiveness of the European information industry (CEC, 1995b). GI2000: Towards a European Policy Framework for Geographic Information. This is an EU policy document intended to raise the level of awareness regarding the development of a European geographic information infrastructure (CEC, 1996b). The key issue addressed in this document is to provide a ‘broad, readily available high quality platform of base data within a uniform infrastructure across Europe so that every market niche is open to every entrepreneur, so that existing data can be combined to provide valuable information’.

A motive for these EU efforts, and similar member state investigations, is to identify dissemination policies which actively promote the greatest openness and innovation in the public and private information sectors. Likewise, it becomes important to identify restrictive policies that may result in reduced opportunities for the successful transfer of raw data sets to commercial data vendors and researchers. The EU has become much more aggressive in promoting better access and efficient utilisation of electronic information. It is equally devoted to enhancing the synergy between the public and private sectors (Bruïne, 1995). For example, three strategic areas identified by a recent INFO2000 initiative are:

¡ exploiting the use of government information; ¡ improving access to non-EC databases; and ¡ the development of the European GI industry. (CEC, 1995b) In addition, there is the explicit objective of strengthening the competitiveness of the European information industry. Perhaps the greatest impediment to utilising publicly

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

165

financed information resources is the controversy over ownership of government resources. The competitiveness of the European information industry or, more specifically, the ability of European geographic information industry to compete in a global setting, will be affected by the issue of how intellectual property rights are applied to public information resources.

Conclusion Comparisons between the EU and/or European member states and the USA are bound to be influenced by the differing legal, political, economic and cultural factors. For example, ‘data access in the US is determined within an ethos of freedom of information whereas in Europe the focus is on data protection and confidentiality’ (Lievesley and Masser, 1993). Despite these differences, European national information policies do share some characteristics with US policies. As noted by legal scholar Henry Perritt, ‘the policy and legal questions on both sides of the Atlantic are remarkably similar. The principal legal questions are whether or not citizens and information resellers have a right of access to public information and, conversely, whether or not the government can block such access by asserting copyright’ (Perritt, 1994, p. 7). The USA does not impose copyright against its citizens in the use of federally produced spatial data, whereas most other nations do. In addition, neighbouring rights protection is viewed by many in the USA as primarily benefiting capital-intensive corporations or large agencies more than small businesses or individual authors. Therefore, this concept is not viewed with the favour currently seen in European nations. These differences are substantial and are likely to be stumbling blocks in the establishment of an international electronic marketplace for spatial data. Other questions as well must be addressed as national governments review their information policies. Should the objectives of dissemination activities be to promote widespread use and access to government information through low-cost access at the cost of dissemination, or should the objective be to generate revenues necessary to fund department activities? Should policy be directed to have the effect of centralising information suppliers or of increasing the diversity of channels and products? Governments must reconcile these sometimes conflicting objectives in a manner which meets the changing information requirements of an information society. Although there are increasing attempts by vested interests in the commercial and government sectors in the USA to move towards more restrictive control over intellectual property rights generally, we predict growing resistance to such attempts by legal scholars, economists, the information industry, the library community and citizen advocacy groups. There appears to be widespread and growing belief in the USA that the general US approach to the treatment of copyright seems more appropriate in the long run for spurring technological innovation and economic vitality for an economy than the alternatives offered by the EU to date.

Acknowledgements This work is based upon work partially supported by the National Center for Geographic Information and Analysis (NCGIA) under NSF grant no. SBR 88–10917. Any opinions, findings and conclusions or recommendations expressed in this material are those of the authors and do not necessarily reflect the views of the National Science Foundation.

166

EUROPEAN GEOGRAPHIC INFORMATION INFRASTRUCTURES

REFERENCES BALLARD, S., JAMES JNR, T.E., ADAMS, T.I., DEVINE, M.D., MALYSA, L.L. and MEO, M. (1989). Innovation through Technical and Scientific Information: Government and Industry Cooperation. New York: Quorum Books. BEASLEY, C. (1995). Tradeable information: a case study of the UK GIS community, Masters thesis, City University, London. BOYLE, J. (1996). Sold Out, New York Times, OP-ED, 31 March, E15. BRANSCOMB, A.W. (1994). Who Owns Information! New York: Harper Collins. BRUÏNE, R.F. (1995). A European Strategy for Geographic Information, Keynote Address Joint European Conference and Exhibition on Geographical Information, The Hague, 27–31 March 1995. BURKERT, H. (1992). The legal framework of public sector information: recent legal policy developments in the EC, Government Publications Review, 19, 483–96. CEC (1993). A Report to the Commission of the European Communities on an Evaluation of the Implementation of the Commission’s Guidelines for Improving the Synergy between the Public and Private Sectors in the Information Market, PUBLAW II Final Report. Luxembourg: DG XIII. CEC (1995a). Commission Green Paper on Copyright and Related Rights in the Information Society, COM(95) 382 Final, 19 July 1995. Brussels: Commission of the European Communities. CEC (1995b). INFO2000, COM(95). Luxembourg: DGXIII. CEC (1996a). Directive on the Legal Protection of Databases, COM(95) 382 Final, Official Journal of the European Communities, 27 March 1996, no. L 77, p. 20. CEC (1996b). GI2000: Towards a European Policy Framework for Geographic Information. Luxembourg: DGXIII). COOPER, P.J. (1986). The Supreme Court, the First Amendment, and Freedom of Information, Public Administration Review, November/December, 622. GOLDSTEIN, P. (1977). Preempted state doctrines, involuntary transfers and compulsory licenses: testing the limits of copyright, UCLA Law Review, 24, 1128–42. GOLDSTEIN, P. (1994). Copyright’s Highway: The Law and Lore of Copyright from Gutenberg tothe Celestial Jukebox. New York: Hill and Wang. HALLO, R. and RODERICK, P. (1995). Freedom of access to information on the environment in the United Kingdom. A user’s guide to the environmental information regulations and EU Directive 90/313. Review of European Community and International Env., 4(2), 220–32. HERNON, P. and McCLURE, C. (1993). Electronic US government information: policy issues and directions, in WILLIAMS, M.E. (Ed.), Annual Review of Information Science and Technology (ARIST), vol. 28, pp. 45–59. Medford: American Society for Information Science. LIEVESLEY, D. and MASSER, I. (1993). Geographic information in Europe: an overview, Conference Proceedings of the Urban and Regional Information Systems Association, 25–29 July 1993, Atlanta, 3, 153–65. LITMAN, J. (1995). Rights in government-generated data, Proceedings from the Conference on Law, Information Policy and Spatial Databases, 28–30 October, 1994, Tempe, AZ, pp. 187–92. University of Maine, Orono. LOPEZ, X. (1993). Database copyright issues in the European GIS community, Government Information Quarterly, 3, 305–18. McCLURE, C., HERNON, P. and RELYEA, H. (1989). United States Government Information Policies: Views and Perspectives. Norwood: Ablex Publishing. NAPA (1993). The Information Government: National Agenda for Improving Government through Information Technology. Washington, DC: National Academy of Public Administration. OFFICE OF THE PRESIDENT (1994a). Establishment of Government Information Locator, OMB Bulletin No. 95–01, 7 December. OFFICE OF THE PRESIDENT (1994b). Co-ordinating geographic data acquisition and access: the national spatial data infrastructure, Executive Order 12906, Washington DC. OMB (1992). Management of Federal Information Sources, OMB Circular A-130, Washington, DC: Office of Management and Budget. ONSRUD, H. and REIS, R. (1995). Law and information policy for spatial databases: a research agenda, Jurimetrics Journal, 35, 377–93. ONSRUD, H.J., JOHNSON, J.P. and WINNECKI, J. (1996). GIS dissemination policy: two surveys and a suggested middle ground approach, Journal of the Urban and Regional Information Systems Association, 8(2), 8–23.

COMPARISON OF EUROPEAN AND US APPROACHES TO IPR

167

PERRITT, H.H., JR (1994). Commercialization of government information: comparisons between the European Union and United States, Internet Research, 4(2), 7–23. PERRITT, H.H., JR (1995). Should local governments sell local spatial databases through state monopolies? Proceedings of the Conference on Law, Information Policy and Spatial Databases, 28–29 October, 1994, Tempe, AZ, pp. 52–72. University of Maine, Orono. POLICY STUDIES INSTITUTE (1995). PUBLAW III, Final Report. London and Namur: Policy Studies Institute GRID. RASKIND, L.J. (1991). Assessing the impact of Feist, University of Dayton Law Review, 17(2), 331–49. SAMUELSON, P. (1996). Consequences of differences in the scope of copyright protection on an international scale. Paper presented at the Forum on Information, National Policies, and International Infrastructure, JFK School of Government, Harvard University, 28–30 January 1996. http://ksgwww.harvard.edu/iip/sampap.html. US CONGRESS (1986). Intellectual Property Rights in an Age of Electronics and Information. Washington, DC: Office of Technology Assessment. US CONGRESS (1990). Helping America Compete. The Role of Federal Scientific and Technical Information. Washington, DC: Office of Technology Assessment. VARIAN, H. (1995). The information economy: how much will two bits be worth in the digital marketplace? Scientific American, September, 200–1.

Statutes and legal cases Berne Convention, 1886. US Treaty Doc. 99–27, KAV 2245. Feist Publications v. Rural Telephone Service Co. 111 S. Ct. 1281, 1287 (1991). Kern River Gas Transmission Corp. v. Coastal Corporation, 889 F.2d 1458 (5th Cir.), cert, den., 498 US 952 (1990). Mason v. Montgomery Data, Inc. (1991 WL 92407 (DD Tex)). NLRB v. Robbins Tire & Rubber Co., 1978. Statute of Anne, c. 19, reprinted in R.S.Brown, copyright 851. Mineola, NY: Foundation Press, 1978. Yeager v. Drug Enforcement Administration, 678 F. 2d 315, 321 (DC Cir., 1982).

Index

access to data 3, 11, 13, 80, 105, 121–2, 131 IPR and dissemination 155, 159–65 legal protection 115, 121–2, 125, 138–40, 142–3, 145 MEDALUS 60, 61, 65 public sector 28–9, 177 adaptations 131 addresses 18 aerial photographs 59–61, 63, 67, 157–8 Agrostat 4 Altai test 158 altitude and elevation 49, 51–2 application programming interfaces (APIs) 91 Association for Geographic Information (AGI) 139 Australia 76, 77 Austria 2, 22, 23, 120, 124 Autodesk 91 availability of data 102–3 BALTEX (Baltic Sea Experiment) 33, 34 Baltic region 2–3, 7–8, 31–42 Baltic Sea 2010 project 2 BDBP (Baltic Drainage Basin Project) 8, 32–3, 35–41 Belarus 34, 37 Belgium 2, 22, 52, 119, 120 geodemographics 72, 76, 77, 82 Berne Convention (1886) 116–17, 128, 142, 157–8 BGIS (Basic Geographic Information of the Baltic Sea Drainage Basin) 32, 38–9, 40–2 branding 88, 92–3 business 22, 71–2, 89, 102, 103 CACI 76, 80 cadastral activities 18 Canada 76, 155 Caroll, Lewis 53 catalogue rights 120, 124, 132 CCN 72, 76, 80, 81, 83 CEN (Comité Européen de Normalisation) 4, 20–1, 42, 93, 119

CERCO (Comité European des Responsables de la Cartographie Officielle) 4, 22 CHR/KHR (International Commission for thé Hydrology of the Rhine Basin) 8, 49, 50–1 Claritas 72, 76 classification of data 72–84 climate 47–50, 58–60, 65, 67 clusters 9–10, 76, 80, 81, 83 compatibility of data 3, 5, 80 competition 11, 17, 28–9, 89, 106, 122–3, 146 IPR and dissemination 156–7, 158, 164 legal protection 116, 120–1, 122–3, 124, 127, 138 computer programs 116–17, 121, 124 confidentiality 138 contract law 138, 158–9 copyright 115–25, 128–9, 145–7, 157–8, 162–3 Baltic region 37, 38, 41–2 data integration 109, 145–7 international aspects 8, 11–13 IPR and dissemination 154–65 legal protection 115–25, 138, 140–51 mapping 127, 128–32 MEDALUS 60–1 pan-European products 26 publishing 89 US federal government 159–62 CORINE (Coordinated Information on the European Environment Programme) 1–2, 104 costs and prices 17–18, 42, 90–1, 97, 104–5, 133–4 data integration 101–6, 109–10, 137, 139, 146–8 international aspects 5, 7–8, 13 IPR 154, 159, 163, 165 mapping 131–2, 133–4 MEDALUS 59, 61 pan-European products 17, 25–9 publishing 87–91, 92, 94–8 transboundary databases 33, 37, 38, 41–2 updating 26, 27, 28 Croatia 22 Cyprus 7, 22 Czech Republic 2, 22, 34, 37

169

170

INDEX

data collection and sources 8–9, 50–3, 106 data integration 101, 104–10, 144 geodemographics 76–80 MEDALUS 58, 59–61 publishing 87–8, 93 Rhine basin 47–9, 50–3, 54 transboundary databases 36–7, 41–2 data integration 2, 5, 9, 12, 101–10 legal protection 137–51 data protection 3, 12, 142, 150 data subject’s rights 143, 144–5 DBMS (database management system) 68 delivery mechanisms 94–5 Denmark 24, 34, 36, 37, 105 copyright 120, 124 SABE 22, 23 derived data 11, 140, 141–2 DGXIII (Directorate General for Telecommunications, Information Markets and the Exploitation of Research) 3, 101, 163 Digital Chart the World (DCW) 36 digitisation 146 dissemination and re-use 71, 110, 150, 153–65 Baltic region 35–8, 41 international aspects 6, 11, 13 IMPACT-2 programme 104, 106, 110 legal protection 115, 119 mapping 130, 132–3 publishing 87–99 drainage patterns 49, 52 Dutch National Council for Geographic Information 6, 115

GDDD (Geographical Data Description Directory) 7, 20–1, 27 genetic information 144 geodemographics 9–10, 71–84 Geolnformation International 88 geospatial information 87 Germany 2, 4, 20, 24, 51, 52 Baltic region 31, 34–8 copyright 119–21, 124, 125 geodemographics 76, 77, 80–3 publishing 90 SABE 22, 23 GI2000 4–6, 17, 41, 42, 164 GILS (Government Information Locator Service) 161 GISCO 34, 37 GMS (geomanagement system) 61–6 government copyright 155, 159–60 Greece 9, 22, 60, 65–7 copyright 119, 120 Greenland 34 GRID (Global Resource Information Database) 8, 104 Baltic region 32, 34–5, 38, 41 GRID-Arendal 34–6, 39 GRID-Geneva 35, 37, 38 Group de l’Interet Economique 19–20

EARS 64 economic activity 17–18 economic or exploitation rights 116 education and libraries 103, 143, 145 ENVIDUCATION 102, 104 environmental issues 2, 4, 7–10, 122, 162 Baltic region 31, 32–42 IMPACT-2 programme 103, 104 MEDALUS 57–68 Rhine basin 47–54 Equifax 76 ERGIS 102 Estonia 34, 36, 37 EUREF 22 EURIPIDES 102, 107 EUROGI (European Umbrella Organisation for Geographic Information) 4, 11, 115, 123, 125 EuroMOSAIC 10, 81–4 Europe 2000 Report 1–2 Europe 2000+Report 1 EUROSTAT 4, 23, 26, 27, 88, 107 Baltic region 34, 37, 38 EXPLORER 102

IACS (Integrated Agricultural Census System) 4 Iceland 7, 22 IGN (France) 9 IMPACT (Information Market Policy Actions Programme) 3 IMPACT Programme Committee (IPC) 101 IMPACT-2 programme 10, 101–10 INFO2000 164 institutional restrictions 137 intellectual property rights (IPR) 10–12, 89–91, 127–35 dissemination and re-use 153–65 IMPACT-2 programme 103, 109, 110 legal protection 115, 116, 149 publishing 88, 89–91, 97, 98 Internet 20–1, 28, 37–8, 54, 153 Baltic region 35, 36, 37–8, 42 MEDALUS 65 publishing 88, 93, 95–9 IRC (International Rhine Commission) 8, 49 Ireland 12, 22 geodemographics 72, 76, 78, 79, 82 legal protection 118–19, 120, 121, 123–5 ISO (International Organisation for Standardisation) 42 Italy 2, 9, 22, 25, 149 copyright 119, 120, 124 geodemographics 76, 77, 80, 82, 83

federal government copyright 159–60 Feist Publications 155–8, 158 Finland 22, 34, 36–9, 76, 120, 124 France 2, 4, 9, 18, 24, 90, 155 geodemographics 76–7, 80–3 MEDALUS 60 Rhine basin 52 SABE 22 Freedom of Information Act 160, 161

Helsinki Commission (RELCOM) 33, 38–9 Helsinki Convention (1974) 7, 32–3 Hungary 22 hydrology and hydrographics 32–4, 134 MEDALUS 58, 63, 64 Rhine basin 8–9, 47–54

Japan 76, 78 JCP (Joint Comprehensive Environmental Action Programme) 33, 39 joint authorship 143

INDEX land cover and land use 4, 9, 20, 25, 27–8, 35, 104 forests 2, 4, 48 MEDALUS 57–68 Rhine basin 47–53 Landsat 67 languages for publishing 92 Latvia 7, 22, 34, 36, 37 Legal Advisory Board 138, 163 legal issues 11–13, 110, 115–26, 127–35, 137–51 IMPACT-2 programme 103, 104, 110 international aspects 5–7, 10, 11–13 transboundary databases 31, 36, 38, 42 lending rights 116, 119, 145 level of achievement 128 liability for data 115, 137, 138 licences and licensing 12, 125, 148–9, 159, 163 integration of data 107, 138, 147–9 publishing 89–92, 98 Liechtenstein 22 Lindop Committee 144–5 Lithuania 34, 36, 37 local government 18, 22, 24–5, 160 IMPACT-2 programme 102, 103, 105, 110 IPR 154, 155, 160, 162 Luxembourg 22, 52 Magill case 123 MAGIS 102, 105 MapBSR 32–4, 39–40, 41, 42 Maplnfo 31 maps and mapping 6, 7, 12 data integration 108, 141, 146–7 IPR 127–35, 155–6, 158, 159, 162 legal protection 118, 120–1, 123 MEDALUS 59–64 overlays 67–8 publishing 87–8 marine science 32–4 market assessment 9–10 IMPACT-2 programme 101, 102, 103, 104–6 publishing 88–9, 92–3, 94–5 marketing rights 132–3 MEDALUS (Mediterranean Desertification and Land Use) 9, 57–68 MEGRIN (Multipurpose European Ground Related Information Network) 4, 7, 19–22, 25–7, 107 Baltic region 40, 41 merged data 129, 140, 157 METEOSAT 64 micro-marketing 71–2, 80, 81 Microsoft 31, 91, 147 missing data 106, 108 monopolies 122, 147–9, 154, 163 moral rights 116 MOSAIC classification 72–3, 76–84 national mapping agencies (NMAs) 20–3, 26–7, 29, 90, 155 Baltic region 36, 37–9, 41 copyright 120–1, 123, 124 Norway 12, 36, 131–2 National Spatial Data Infrastructure (NSDI) 5, 161–2 national statistical agencies (NSAs) 37 National Statistical Institute 23

171

neighbourhood 72–5, 81–2 neighbouring rights 116, 132, 158, 163, 165 Netherlands 2, 4, 22 copyright 119–21, 124, 125 geodemographics 72, 76, 78, 80–2 Rhine basin 51, 52 Netherlands Council for Geographic Information 6, 115 Netherlands Institute for Environmental Research (RTVM) 50 New Zealand 76, 78 Nordic Cartographic Database (NCD) 36, 37, 41 Norway 22 geodemographics 76, 79, 82 IPR 131, 132 transboundary databases 34–7 Norwegian Mapping Authority (NMA) 12, 36, 131–2 NUTS (Eurostat Nomenclature des Unites Territoriales Statistiques) 18, 22, 23, 36, 107 Office of Management and Budget (OMB) 161 OGIS (Open GIS Consortium) 6 OMEGA 102, 108 Ordnance Survey 88, 139, 146 ownership 10–11, 26, 89–90, 106, 110, 140–3 copyright 119, 124–5, 130, 132 data integration 139, 140–51 IPR 158, 165 use restrictions 150 packaging 88, 92–3, 94–5, 97, 98 Paperwork Reduction Act 160–1 patents 147–8 PETIT 20 photographs 120, 124, 125, 144, 155 aerial 59–61, 63, 67, 157–8 planning 2, 8, 33, 35, 72 Poland 34, 35, 36, 37 pollution 32–5, 50 PONET (Political and Oceans Layer) 36 population 35, 36, 37, 38 Portugal 9, 22, 120, 124 privacy 115, 138, 140, 144–5 public sector 2, 17–18, 22, 24–5, 28–9, 160 copyright 121–3, 125 IMPACT-2 programme 102, 103, 105, 110 IPR and dissemination 154, 155, 160–5 legal protection 137, 138, 140 publishing 2, 10, 87–99 queries in MEDALUS 68 rasters 37–8, 40, 51 RAVI 115 records of fact 140–3, 147–8, 149 re-use of data see dissemination and re-use Rhine river basin 2–3, 8–9, 47–54 RHINEFLOW 8, 47–53 RIVM 50 roads 2, 27–8, 42 royalties 90, 91, 95, 97, 148 data integration 105, 107, 109, 110 runoff 47–9, 58, 60, 65, 67 Russia 34, 37

172 SABE (Seamless Administrative Boundaries of Europe) 7, 20–2, 25–7, 41, 107 sales channels 94–9 SCOLE (Seamless Coastline of Europe Database) 22, 25 screen displays 131 Slovakia 22, 34, 37 Slovenia 22 socio-economics in MEDALUS 65 software front end 88–9, 90, 91–2, 97 soil types 47–53, 60, 63, 65–7 South Africa 76, 79 Spain 2, 9, 22, 51, 65–7, 120, 124 geodemographics 76, 79, 81, 82 Spot 67 standards and standardisation 1, 4–5, 9, 20–1, 106–7, 119 data integration 105, 106–7 geodemographics 71–84 mapping 134 MEDALUS 59, 61 publishing 93 transboundary databases 39, 40, 42 Sweden 22, 25, 34–9 copyright 120, 124 geodemographics 76, 79, 80, 82 Switzerland 7, 22, 23, 36, 51, 52 synergy 5, 163, 164 third party rights 143–4 timetables 105, 109 TiTAN 102, 108 topology and topographics 17, 19–20, 22–3, 26–8, 87

INDEX tourism 102, 103, 109 transboundary databases 2–3, 31–42, 71–84, 107–8 tribunal 149–50 Ukraine 34, 37 UNEP (United Nations Environment Programme) 34–6, 38, 104 United Kingdom 4, 12, 22, 35 data integration 137–41, 144, 146–51 geodemographics 72–3, 76–8, 80–2 IPR 131, 154, 155 legal protection 118–19, 120–1, 123–5 publishing 88, 90, 91 United States of America 5, 12–13, 88, 144, 153–65 copyright 146, 159–60 geodemographics 72, 76, 79, 80, 81 United States Federal Geographic Data Committee 5 Universal Copyright Convention (1952) 116 updating 12, 21, 26, 41, 93, 118, 124 costs 26, 27, 28 data integration 104–5, 108–10 utilities 18, 87 value added data 139–40 VASAB (Vision and Strategies for the Baltic Region) 2010 Project 33, 39 vectors and vectorisation 38, 40, 121 VITAL 102, 108 water quality 31, 32, 34, 47, 49 Younger Committee 145