INFRASTRUCTURE FOR SPATIAL INFORMATION IN
EUROPE (INSPIRE): FROM CARTOGRAPHY TO SPATIAL
OBJECTS AND NETWORK SERVICES
Katalin. Tóth and Paul. Smits
European Commission, DG Joint Research Centre
“Scientia potentia est”1. Although the famous aphorism was formulated only in the
XVI century, the mankind since the beginning of its socialisation tried to share knowledge
to increase its potential for surviving and progress. Sheer facts, however accurate they are
stand-alone pieces of information never achieve the same effect that they do when they are
put in context. Time and space are one of the most frequently used references. If we know
what is existing or what is happening, we are only halfway to the knowledge. The
questions who, what, where, when, and perhaps many others have to be answered in order
to complete our knowledge and turn it into useful information.
The spatial context, following the classification of Bregt (2004), can be described by
three distinct frameworks. The first is the geodetic framework, which is the determination
of the Earth’s size, shape, and locations. This framework started with works of
Eratosthenes and Ptolemy , and although it still continues developing it reached its apex in
the sixteenth and seventeenth century.
The second framework is the topographic framework. It started in 1669 in France,
when Colbert, upon an order of King Louise XIV, asked Cassini to create a topographic
map of France. This initiative was followed by the development of the institutional
framework by establishing national topographic surveys and mapping agencies.
The third is starting nowadays, and can be called the geo-informatics framework. It
concerns organising the integration of spatial data and the access to this information
creating spatial data infrastructures.
Each framework constitutes the foundation for the next one. It is interesting to see that
the start of each of them is accompanied by considerable political attention, and that the
players change (geodetics, surveyors, geo-informatics) (Bregt, 2004).
1 “The knowledge is power” - aphorism of Sir Francis Bacon (1561-1626)
The portrayal of spatial context already appeared in the Sumerian and Babylonian clay
tablets and have been present in each framework. However the development of scientific
cartography roots in the antic Greece, culminating in the “Geograpia” of Ptolemy
[Klinghammer 2005]. Mapping as powerful tool for communicating information became
integral part of geography, astronomy and geodesy and reached step by step many other
disciplines. This process triggered the constitution of the independent cartographic science,
which took pace according to Klinghammer between 1869 and 1925.
The revolution in information technology has challenged cartography to streamline not
only the technology and production lines, but also to adopt appropriate changes in the
paradigm. From focusing on the process of creating high quality maps with clear visual
expression the centre of interest shifted to the process of creating cartographic information,
converging with other disciplines involved in development of Geographic Information
Systems. Cartography embedded in information technology is dealing now with objects,
spatial modelling and services often transmitted via the web. Parallel to the penetration of
GIS in various application fields and the every day life the need for integrating the existing
stand-alone system has established the spatial data infrastructures.
The aim of this paper is to provide some insight in the role of cartography in Spatial
Data Infrastructures, with a particular focus on the European Spatial Data Infrastructure as
it is envisioned by the Infrastructure for Spatial Information in Europe (INSPIRE)
Directive. It does so by highlighting some of the experience we obtained in the preparatory
work for INSPIRE, mainly from organisational, data, and service views.
Cartography and Spatial Data Infrastructures
In Spatial Data Infrastructures the single purpose data collection should be replaced by
information management to facilitate the access to and sharing of information in order to
avoid duplications in data collection and processing. For effective information integration
in GI a common regulatory and technical framework is needed, laying down the principles
and providing guidance how to integrate and share business and geographic information. In
addition to the economical benefits a single and agreed set of reference data used for
geographic positioning of thematic information supports interoperability, data integrity and
The importance of cartography in creating the reference framework is evident; many of
its elements have been established by the cartographic science. Cartography has valuable
achievements how to abstract real world in conceptual models both in case of reference
and thematic data, how to move across different levels of details. Visualisation continues
being important, as information should be communicated to the users in an appropriate
way. 2-D media like screen and paper are still the main channels of visualisation, therefore
the projection systems are indispensable contributions of cartography to SDIs. As
suggested by the three frameworks in the introduction, the cartographic science forms
integral part of spatial data infrastructures, directly contributing to data harmonisation and
ESDI as shaped by the INSPIRE Directive
In the European Union environmental policy making is an important driver for
establishing the SDI. High level requirements from the environmental policy makers, like
better information needed to support policies, the improvement of existing information
flows, appropriate consideration of diversity across regions, the need to revise the approach
to reporting and monitoring, and moving to the concept of sharing and integration of
information, have met with a number of obstacles. These obstacles range from data policy
restrictions, lack of co-ordination across borders and between levels of government, till
lack of standards leading to incompatible information and information systems,
fragmentation, information, redundancy and impossibility of data reuse.
The INSPIRE Directive of the European Commission , agreed now by the European
Parliament and the Council2, is aiming at addressing the aforementioned obstacles. The
European SDI that INSPIRE envisions will be built upon interoperable SDIs that Member
States are requested to establish and operate.
The infrastructure will be set up through 5 components: metadata, harmonised data
specifications, network services, agreements for data sharing, and measures for
The European SDI must be based on the infrastructures already in place or in the phase
of development in the Member States of the European Union. In order to bring the
initiative to success the provisions for the implementation should be based on the
consensus of the stakeholders who range from data users through data transformers and
SDI coordinators till data and information providers.
For each INSPIRE component a specific working group has been created by the
European Commission to elaborate implementation provisions through specifications.
2 The joint text was adopted by the Council on the 29/01/2007 and the Euroepan Parliament on 12/02/2007 It
is expected to enter in force in April/May 2007.
From point of view of cartographic science specifications for harmonised data and network
services are the most relevant building blocks, which will be discussed in more details in
the following paragraphs.
Data harmonisation in INSPIRE
Data has crucial role in information infrastructures. Data collection is its most
expensive part that can account for around 60-80% of the total costs of setting up a
system3, therefore data reuse is of prime economic interest. Good metadata pave the road
for this, however the reuse depends on how much data is comparable with the actual need
in terms of semantics, schema, format and data matching. Some of these obstacles, for
example differences is the format can be easily solved by conversion, while others, like
schema mapping need more or less manual work and require comparable conceptual
In a service oriented architecture the generic aim is to provide interoperability within
the components of the system using such services that guarantee flexibility by on-demand
operations. On the other hand optimal use of services requires harmonisation efforts in
semantics and schema level too. The INSPIRE Directive foresees harmonisation in 34 data
themes that are necessary to build effective environmental information systems (table 1).
The data themes were devised in three annexes according to their roles in the SDI. The
majority of data necessary for referencing environmental information is included in Annex
I and II, while thematic data is present in Annex III. This grouping also defines the
roadmap of harmonisation efforts. Implementing rules for Annex I theme have to be ready
within 2 years from the time when the Directive enters in force, while for Annex II and III
the targeted timescale is 5 years. After 7 years from adopting the implementing rules all
electronic data within the scope of the Directive and owned by public organisations, or
third parties on their behalf, have to comply with the specifications laid down in the
The other aspect of classifying themes in annexes deals with the degree of
harmonisation. According to Article 7 (4) geo-referencing, definition and classification of
spatial objects must be specified for each theme. In addition, as required in Article 8(2), for
Annex I and II the following aspects must be also considered:
o common system of unique identifiers for spatial objects;
3 Source: PANEL-GI Compendium, EUR 19360en, 2000
o relationship between spatial objects;
o key attributes and corresponding multilingual thesauri;
o provisions for the exchange of the temporal dimension;
o provisions for the exchange of updates.
Annex I Annex II
1. Coordinate reference systems
2. Geographical grid systems
3. Geographical names
4. Administrative units
6. Cadastral parcels
7. Transport networks
9. Protected sites
11. Land cover
14. Statistical units
17. Land use
18. Human health and safety
19. Utility and governmental services
20. Environmental monitoring facilities
21. Production and industrial facilities
22. Agricultural and aquaculture facilities
23. Population distribution – demography
24. Area management/restriction / regulation
zones & reporting units
25. Natural risk zones
26. Atmospheric conditions
27. Meteorological geographical features
28. Oceanographic geographical features
29. Sea regions
30. Bio-geographical regions
31. Habitats and biotopes
32. Species distribution
33. Energy Resources
34. Mineral resources
Table 1 - Data themes of INSPIRE as listed in the annexes of the Directive
It is very important to clarify from implementation point of view when data can be
regarded harmonised. The current proposal is that in INSPIRE, harmonisation efforts will
be done mainly at application schema level, but some operational measures (registers,
terminology dictionaries, object classification, multi-language definitions, system for
unique identifications) are required as well. They may be completed by guidelines and best
practice descriptions to support consistent implementation of the specifications. The
proposed components of harmonisation are shown in table 2 that are that are addressed in
the INSPIRE Generic Conceptual Model.
(C) Reference model
(H) Object referencing
(G) Coordinate refe-
rencing and units model
(A) INSPIRE Principles (B) Terminology
(D) Rules for application
Schemas and feature
Management (L) Registers and
(I) Data translation
(E) Spatial and temporal
aspects (F) Multi-lingual text and
(M) Metadata (N) Maintenance (O) Quality
(P) Data Transfer (R) Multiple
(J) Portrayal model
(S) Data capturing
Table 2 – Data harmonisation components
The Generic Conceptual Model is the basis for the development of implementing rules
that will be expressed through data product specifications for each theme. There are
normative elements in the Generic Conceptual Model; geometrical, topological and
temporal representations together with the system for unique identification of spatial
objects and reference system must be strictly followed.  Only additional concepts of the
thematic area must be modelled as part of the application schema for the data theme.
However thematic application schemas may have dependencies between each other. These
dependencies stem from cross thematic consistency requirements that should be addressed
during the application modelling.
Spatial cognition and modelling
Correctness of information is of utmost importance for the users. It depends not only
on the quality of data per se stored in the systems, but also on how the data describe reality.
Geographic reality cannot be measured exhaustively because it is nearly impossible to
obtain measurements for every point across the entire landscape Therefore, a fundamental
discrepancy exists between geographic data and the reality that they are intended to
represent. This discrepancy or uncertainty is propagated through and may be further
amplified by data management and analyses [Tóth and De Lima 2005]. Good starting point
that means appropriate modelling has determining role in the quality of geographic
ISO TC 211 proposes a straightforward way for application schema development.
Depending on the actual interest the Universe of discourse is modelled in meta model with
feature types that are later represented by the appropriate elements of the spatial schema.
The feature types can be either genuine concepts to be documented in the feature catalogue
or can be equally taken from an existing one. From SDI point of view this second has to be
followed whenever possible, which leads overall consistency in the infrastructure. The
classification of the real world objects should be done without overlaps and gaps; feature
and attribute definitions must be unambiguous.
The ISO 19109:2005EN standard, being high level and generic cannot consider the key
point of correct modelling: the selection rules, which of course depend on the users’ need.
The semantic resolution, or “scale of reasoning” [Ruas, 2003] defines the granularity (level
of details) of the model. The same object may be essential in one application, while
worthless in another. Similarly the same phenomenon may appear at object or aggregation
level, with simpler or more complete attribute set.
Meta levelApplication levelData level Meta-meta level
Figure 1 – Modelling process as proposed in ISO 19109:2005 EN completed with the role
of the level of details
The level of details defines how to describe spatial objects4 in terms of the spatial
schema and the “acceptable” quality that in their turn define the applicable data collection
methods as well. Therefore the whole modelling - system implementation – data collection
cycle should be governed by the targeted level of details, as shown in fig.1.
4 In the INSPIRE Directive the term “spatial object” is used insteaad of “feature”.
View 2 View 1
It is not difficult to see that there is no theoretical limit for increasing the semantic
resolution. Never the less each model has its own limit. This limit usually is connected to
spatial resolution level, however spatial resolution is not the only classifier in this process
(e.g. landmarks – like churches, single trees, etc. - are not deselected, reclassified or
aggregated). Due to the long traditions and experience cartographers may successfully
describe the interdependency of spatial and semantic resolutions, or formalise the
contextual classifiers of objects.
Multiple Representation and Data Consistency
The same real world object depending on the intended use can be modelled from
different views (thematic views) and within each view with different level of details
(fig.2). Multiple-representation, both intentional and unintentional, is an everyday practice
in geographic information technology that should be also addressed within the frame of the
SDIs in order to support decision making and the reuse of the data. People need to view
information coming from different sources and at different resolutions in a consistent way.
Figure 2 – Objects modelled from different views (A) and in different spatial resolution (B)
In this respect Article 13 of the INSPIRE Directive states: “The implementing rules
shall be designed to ensure consistency as between items of information which refer to the
same location or between items of information which refer to the same object represented
at different scales.”
The above paragraph requires that the INSPIRE implementing rules address the issues
of multiple representation and data consistency. The problem of data consistency in SDIs is
raised when combining data from different sources. Obvious consistency constrains occur
for example in case of digital elevation models and hidrography, cadastral parcels and
administrative units, addresses and buildings, etc. These and other related datasets are
being identified during the INSPIRE theme definition and scoping. For developing
INSPIRE Implementing Rules themes with strong inter-relations have to be grouped and
specified together. It must be explicitly stated that two datasets are comparable in terms of
consistency and the specific rules should be developed and formalised using constraints.
The actual level of details of the data themes will be defined based on the analysis of
the user requirements. Spatial Interest Communities are invited in the process not only
from the part of data providers, but also from users. Based on selected Community policies
it can be predicted that the required scales/Levels of Detail (LoDs) may span from large
(noise maps at street level according to the Directive 2002/49/EC) to small (spatial
representation of the river basins according to Directive 2000/60/EC). Recent SDI studies
show on one hand that a selected data theme may be available at different LoD from
country to county; on the other hand a data theme can be covered by diverse datasets even
within a country that should be considered when data are to be harmonised.
For a single data theme depending on the LoD several application schemas can be
developed, however this must be justified by appropriate use-cases. Basic principle should
be to use as few LoD as possible, which simplifies not only the data modelling, but also
organisational and management aspects of maintenance.
The INSPIRE data specifications presumably will be governed by ISO 19131 (Data
Product Specifications), which sets requirements amongst others to the data structure and
accuracy that are closely related to the level of details. The data consistency can be one of
the aspects of “spatial data validity” to be documented in the metadata (Article 5,
As a piece of European legislation a directive sets out the objectives to be achieved,
while does not regulate how to do so. Member States are free to decide how to implement
the specifications of INSPIRE. Methods, implementation technologies and procedures are
out of the scope of the Implementing Rules; however they can be made available as best
practice examples. Such best practice examples are generalisation, as a method for creation
of multiple-representations across the scales, and schema integration supported by
Consistency requirements change with the scale, thus generalisation gives a reasonable
framework for their modelling. Taxonomic and partonomic relations that are expressed in
metric, topological and longitudinal terms give quantitative measures for consistency too.
The experience accumulated in cartographic science in this field gives another valuable
contribution for building SDIs.
Sound data integrity should be coupled with visual coherence at all resolutions. In the
traditional paper based cartography data collection, editing and portrayal formed a unique
technological process with high human involvement and continuous quality control. This
has been changed by the digital technology, where users can easily customise the
presentation and the layout of information. This flexibility however may create
uncertainties especially among those communities that had established standards for
legends, symbols and other graphical elements. The conventions of standard portrayal have
to be respected in the SDI, therefore appropriate arrangements and tools should be put in
place. Standard portrayal rules can be specified at data level as default values, which can
be later modified by the user.
There where the above mentioned flexibility is needed, the default portrayal rules can
be replaced by those that are deemed more appropriate for a given application. Web
services that instantiate recent versions the Web Mapping Service interface specification
are a good example of how a SDI provides this level of flexibility.
Adaptive zooming underpinned by multiple-representation, appropriate object
referencing and generalisation capabilities create “seamless” viewing perception of users
when information is aggregated of disaggregated. This relatively new feature of
visualisation adds value as compared to the traditional collection of maps sheets or
A related topic is automatic generalisation. The authors do not consider automatic
generalisation methods to be mature enough to be considered as a service component in the
ESDI. For practical reasons multiple scale representation is always needed, which can be
completed by generalisation . For static application schemas the on-the-fly
generalisation can be replaced by database views.
Another challenging task in modern information management is managing and
communicating data quality. This is especially relevant in case of WFS, where spatial
objects can come from everywhere theoretically, bringing with them all the uncertainty of
their whole production chains. Naturally the users do not want to waste too much time on
reading metadata for each element; they need a simple and quick way for communication.
Quality visualisation offers an effective “on the spot” judgement about the “fitness for use”
and data uncertainties.
The three frameworks mentioned in the introduction, namely the geodetic, the
topographic, and the geo-informatics frameworks, illustrate that cartography is an
important aspect of the SDI concept, and cannot be de-coupled from SDI development.
Rather, the relevant expertise must be channelled into SDI development processes.
The experience with the process of data specification work within INSPIRE bears
witness to this. Describing and classifying reality with consistent models, the spatial
cognition, and communicating information according to thematic needs of the users, are
only few out of many examples of expertise that the cartographic community can provide
to the process of establishing a SDI.
It is therefore paramount that the cartographic community recognises the event of
upcoming SDIs as an enormous opportunity to connect this fascinating field to the other
disciplines that consider themselves to be part of the SDI community. One way to
contribute to this, is to make the knowledge and expertise available by participating in
standardization initiatives of OGC, ISO/TC 211, CEN/TC 287, and INSPIRE.
 Directive of the European Parliament and of the Council establishing an Infrastructure for
Spatial Information in the European Community (INSPIRE)
 INSPIRE Generic Conceptual Model (2007) Manuscript to be published in http://inspire.jrc.it/
 ISO 19109:2005EN Geographic Information - Rules for application schema
 Recommendations of the INSPIRE Multiple-representation and Data Consistency Workshop,
Ispra, 7-8 November, 2006. http://sdi.jrc.it/ws/multiple_rep/links.cfm
Bregt, A. (2004)
Het derde raamwerk (English: The third framework). GEO-INFO 2004-11. http://www.geo-info.nl
Klinghammer, I (2005)
A térképészet tudománya. Akadémiai székfoglaló előadás 2005 február 15. (The Cartographic Science)
Membership inauguration speech of 15 Februry 2005 at the Hungarian Academy of Science, Budapest.
Ruas, A. (2003)
Spatial Analysis and Agent modelling to automate generalisation process http://geopro.cic.ipn.mx/Pruas.pdf
Tóth, K., Nunes de Lima, V. (2005)
Data Quality and Scale in Context of European Spatial Data Harmonisation (11th EC-GIS workshop,