Finnish Environment Institute | Suomen ympäristökeskus | Finlands miljöcentral

sykefi header valkoinen

Producing land cover and land use data in CORINE Land Cover 2000 and Image 2000 projects in Finland

+ Show information

Basic project information

A satellite image mosaic and both raster and vector land cover databases of Finland were produced by Finnish Environment Institute (SYKE) as a part of the European CORINE2000 Land Cover and IMAGE 2000 projects (CLC2000, IMAGE2000). CORINE2000 classification of Finland is based on automated interpretation of satellite images and data integration with existing spatial datasets.

Background and Objects

European Environment Agency (EEA) and Joint Research Centre (JRC) launched the Europeanwide IMAGE2000 and CLC2000 projects in order to identify main land cover changes in Europe.  Corresponding project (CORINE Land Cover 1990) was executed in the beginning of 1990's when land use and land cover of  the contemporary European Union countries were mapped first time using satellite images, common classification criteria and quality standards.

The majority of the participating countries produced CLC2000 database by updating  the CLC1990 database using the traditional method of visual interpretation of the satellite images. In addition to CLC1990 and CLC2000 databases also the land cover changes (1990-2000) database was produced.

In Finland the primary objective in CLC2000-project was to produce a accurate raster land cover database for national use. Vector database for European use with quality standards set by EEA was generalized from the national raster database. Databases are based on automated interpretation of satellite images and data integration with existing digital map data. Finland was not involved in CLC1990 project and don't have a CLC1990 database or the land cover changes database.

Data and methods

Satellite image interpretation

CORINE2000 is based on LANDSAT 7 ETM satellite images (years 1999-2002) from which IMAGE2000 national satellite image mosaic was produced. IMAGE2000 and field data were used in estimation of  continuous variables describing vegetation type and coverage and to update the existing map data.

The selected satellite images were geometrically corrected by Metria Sweden. Radiometric preprocessing included atmospheric correction and topographic correction in Northern Finland. The mosaicking of individual satellite images were carried out in order to get stratumwise and nationwide mosaics for interpretation and visualization purposes.

In forests and semi-natural areas as well as in wetlands following land cover variables were estimated from the satellite images: tree height (m), tree crown cover (%), volume of broadleaved trees (m3/ha) and total volume (m3/ha). Additionally in northern Finland ground vegetation type was estimated for non-forested areas.  Continuous land cover variables were transformed into discrete CORINE classes by thresholding the variables according to the class descriptions in CORINE nomenclature.

SYKE and the Technical Research Centre of Finland (VTT) developed a production line based on calibration of satellite measurements and automated satellite image interpretation. Metsähallitus and UPM provided field datasets.

Semi-automated approach was used in updating of map data and in mapping of coastal and inland marshes.

Use of GIS-data and data integration

Satellite image derived land cover data were combined with existing digital land use and soil information. The principal geographic data sources were SLICES- land use database, soil information from the Topographic Database of Finland produced by the National Land Survey of Finland and the Population Information System produced by Population Register Centre.

A national raster database with pixel size 25 by 25 meters was produced first. A vector  database accordant with CLC2000 definitions was generalized from the national raster database. An automated generalization procedure was developed in order to produce required vector data from 25 meter raster data. In vector database the minimum mapping unit is 25 hectares and minimum width of units is 100 m.

Results

The main outputs of the project are a national satellite image mosaic (IMAGE2000), raster land cover database with a resolution of 25 m times 25 m for national use (Fi25m) and CLC data in vector format with minimum mapping unit of 25 hectares for European use (EU25ha).

IMAGE2000 satellite image mosaic & CLC2000 vector database (EU25ha) have been combined in EEA to form a European wide digital image mosaic and vector database.

Classification

The CORINE nomenclature is hierarchical and it has five classes in level 1: artificial surfaces, agricultural areas, forest and seminatural areas, wetlands and water bodies. These are subdivided to 15 classes in level 2 and to 44 classes in level 3. Not all the classes (e.g. olive  groves and rice fields) exist in Finland.

In Fi25m –product there are  also classes in national level 4, for example forest classes have been divided to subclasses by soil information.

Estimation of  accuracy

The accuracy and the quality of the results were estimated by Finnish Forest Research Institute and by EEA. The evaluation of the accuracy of thematic content has been estimated during the interpretation process by project personnel.

The Finnish Forest Research Institute validated the results by comparing the interpretation results to National Forest Inventory –information (NFI). When Finnish national CLC2000 –classification is compared to classification based on NFI-information, the overall accuracy of main classes is about 90%. The overall accuracy at the second classification level is about 81% and at the third classification level about 70%. It should be remembered that the CLC2000 and NFI-classifications are bit different, especially concerning artificial surfaces and agricultural land.

The accuracy and compatibility of the 25 ha vector databases produced by participating countries (EU25ha) has been evaluated by EEA.  Of  the four per cent of the area of Finland pattern borders and classification were checked for the evaluation purposes.

According to the reliability checking done for  the satellite images by SYKE, the accuracy of georeferencing of satellite images was quite good, the average planimeric RMSE for all images was 12.9 m. RMSEs are 6.1. – 18.9 m, depending on image. The determination of RMSE was based on independent test ground points.

Applications

The information collected and methods developed during the project can be further used in different monitoring and research projects concerning land cover and land use.

In Finland the information will be firstly used in the monitoring of the land use and the natural resources and the research of the biodiversity. The results will be utilized as input in various environmental models, for example in flood prediction and in estimation of the nutrient leaching into water bodies. Information is also needed when implementing various directives e.g. European Union's Water Framework Directive.

In European Union level information is used when assessing the effects of Community's agriculture, regional or environment policies.

Funding and co-operation partners

The CLC2000 project was co-funded by the EU and Finland. Nationally the project was funded by SYKE, the Ministry of the Environment and the Ministry of Agriculture and Forestry. The EU granted half of the funding and also supervised the project. SYKE is responsible of the national cooperation with the EEA.

SYKE has been responsible for the data production in Finland. SYKE worked in co-ordination with Remote Sensing group in Technical Research Center of Finland (VTT Information Technology), Metsähallitus, UPM and Metla.

Further information

Finnish Environment Institute (SYKE),
Data and Information Centre, Geoinformatics Division
P.O. Box 140, 00251 Helsinki

Pekka Härmä, tel. +358 400 148738, firstname.lastname@ymparisto.fi [pekka harma]

Related publications

Härmä, P., Teiniranta, R., Törmä, M., Repo, R., Järvenpää, E., Kallio, M. The Production of Finnish CORINE Land Cover 2000 Classification.   XXth ISPRS Congress, Istanbul, Turkey, 12.-23.7.2004. International Society for Photogrammetry and Remote Sensing

Törmä, M, Härmä, P. Accuracy of CORINE Land Cover Classification in Northern Finland.   International Geoscience and Remote Sensing Symposium, 20-24 September 2004 Anchorage USA

Härmä, P., Teiniranta, R., Törmä, M., Repo, R., Järvenpää, E., Kallio, M. Finnish CORINE Land Cover 2000 Classification.   International Geoscience and Remote Sensing Symposium, 20-24 September 2004 Anchorage USA

Törmä, M., Härmä, P.  Topographic Correction of Landsat ETM-Images in Finnish Lapland.  International Geoscience and Remote Sensing Symposium, July 21-25 2003 Toulouse France.

Final Report

Published 2013-04-26 at 13:15, updated 2013-05-07 at 10:39

Target group: