in MS Access. In 1998 it has been decided to 
enlarge the scope of the meta database to regional 
systems and to the accession countries. 
In addition to that, meta data on information systems 
on landscapes will also be introduced. For that 
purpose the database structure will need to be 
reviewed and improved. Access for the wider public 
via the Internet will be subject of discussion with 
Eurostat that is the owner of the database. 
2.2 Methodological studies 
Studies on statistical methodology contribute to the 
evolution and the development of the European 
statistical system. In this respect, statistical methods 
for the use of new tools in statistics, such as remote 
sensing and geographic information system, and for 
harmonisation of data from the Member States 
regarding land use information are subjects of 
studies. 
2.2.1 The Statistical framework for Remote 
Sensing Applications in Statistics 
The use of remote sensing imagery as a source of 
geographic and statistical information requires 
adopting a work programme that respects a 
coherent statistical scheme. The following 
components are required: a statistical nomenclature 
of land cover and use, clear guidelines and 
specifications for fields of observation, units of 
observation, methods of observation and methods 
of classification. 
CESD LU has developed such a framework for the 
account of Eurostat. This framework served in pilot 
projects as part of the technical Terms of Reference 
for the contractors carrying out these projects. It 
includes a pilot nomenclature entitled CLUSTERS 
(Classification for Land Use STatistics - Eurostat 
Remote Sensing programme). The nomenclature 
respects the principles of thematic coherence (one 
class of the classification system corresponds to 
one type of land use and one type of land use can 
only be put into a single class, no mixed classes), 
consistency of scale (the user must be able to 
aggregate or desegregate results according to the 
required analysis - work on different scales / 
restitution units), spatial coherence (compatibility 
over all the area of interest) and consistency over 
time (taking stocks, not flows to enable change 
detection over time). 
2.2.2 Study on a reference land use 
classification 
In the framework of this study, carried out for the 
account of the FAO, the concepts and the existing 
approaches of land use information systems have 
been described followed by a preliminary design of 
a reference land use classification system 
containing land use classes defined by the 
combination of a set of independent diagnostic 
attributes, so-called classifiers. The study 
recommends to develop a complete land use 
classification based on the functional approach 
(socio-economic purpose) and to complete the 
already established 'Land Cover Classification 
System' of the FAO (mapping land cover based on 
EO data) with a sampling approach on land use. 
2.2.3 Indicators on Landscape fragmentation in 
Europe 
The potential and the limitations of the existing data 
may be analysed best through the use of the data. 
The study, carried out in the framework of a 
collaboration between Eurostat, DG AGRI and the 
DG JRC and the EEA, exploits the only existing 
geographic land cover / land use database covering 
(nearly) all Europe, the CORINE Land Cover 
database, for the calculation of fragmentation 
indices to analyse the structure of landscapes in 
Europe. Among the different technical points 
discussed (definition, comparison and use of 
indexes) the experiment raises the problem of the 
choice of the restitution units and its effect on results 
(NUTS regions, ‘natural’ landscape units like 
‘Naturraume’ in Germany) while a solution based on 
a regular moving window grid is tested. Beyond the 
results, limitations of CORINE Land Cover data are 
stressed and particularly those concerning the 
definition of the nomenclature (mixed classes due to 
both, the size of the minimum mapping unit and the 
existence of intrinsically mixed areas on the earth 
surface) and the temporal heterogeneity of the 
satellite images used to build the database. 
Because of their impacts on results, it appears as 
obvious that the definition and the elaboration of a 
homogeneous and coherent database at European 
level are of importance. 
2.3 Pilot projects 
Pilot project activities involving the application of 
earth observation data for statistical purposes help 
to demonstrate the potential of this type of 
information source but also its limitations. CESD LU 
managed pilot projects for the account of Eurostat, 
national statistical services and local authorities that 
means users of the data. The technical work was 
carried out in most cases by the private sector. The 
subjects of the projects varied from purposes in 
agriculture or urban statistics to regional planning 
and environmental monitoring. 
The following map provides an overview of the 
project sites. 
The applications for different purposes were always 
based on the retrieval of land cover and land use 
information using satellite imagery and other 
existing data (such as Digital Terrain Models, digital 
maps e.g. on infrastructure, land use planning etc.) 
according to the statistical framework mentioned