Ecosystem coverage

Indicator definition

Proportional and absolute change in extent and turnover of land cover categories aggregated to relate to main ecosystem types in Europe from 2000 to 2006.

The 12 ecosystem types discussed represent (1) urban, (2) cropland, (3) agricultural mosaics, (4) woodland and forest, (5) grassland and tall forb, (6) heathland, shrub and tundra, (7) transitional woodland, (8) sparsely vegetated land, (9) inland wetlands, (10) coastal, (11) rivers and lakes and (12) marine areas. This indicator is based on photo-interpretation of satellite imagery, and gives a 'wall to wall' picture of the changes and dynamics in Europe with respect to ecosystems. Additional indicators can be used to further highlight trends in extent and state of each of the ecosystem types mentioned above using computations from other data sources.

Units

Land cover classes area change in ha

% change of land cover classes coverage

Policy context and targets

Context description

This indicator uses photo-interpretation of satellite imagery to give a rough picture of the trend in the observed area and proportion of the major ecosystems in Europe since 1990.
Satellite imagery offers the potential to characterise land cover over very large areas efficiently and very cost effectively. It is possible to produce land cover maps from satellite imagery based on the spectral properties of each pixel within a scene. By grouping pixels into classes with similar spectral properties and associating these classes with particular land cover types, it is possible to produce maps which delineate land cover. Land cover change is then used to indicate the trends in the extent of major ecosystems, such as forests, croplands, wetlands, etc. For this indicator we use data from the Corine land cover database (CooRdinate Information on the Environment - Corine).

 The CLC data are based on 44 land cover classes that are aggregated into 12 ecosystem types for the purpose of this indicator. Spectral properties allow the CLC project to distinguish between land cover classes. For example, CLC has three classes showing forest land cover: broad-leaved forest, coniferous forest, and mixed forest. By aggregating the information of these three land cover classes we have information on the extent of the forest ecosystem within the limitations of the CLC data (see section on main disadvantages). The CLC data however are the best available at present to cover large areas of Europe in a harmonised way.

Targets

EU Biodiversity Strategy 2020 - headline target and Target 2

Related policy documents

• EU 2020 Biodiversity Strategy
  in the Communication: Our life insurance, our natural capital: an EU biodiversity strategy to 2020 (COM(2011) 244) the European Commission has adopted a new strategy to halt the loss of biodiversity and ecosystem services in the EU by 2020. There are six main targets, and 20 actions to help Europe reach its goal. The six targets cover: - Full implementation of EU nature legislation to protect biodiversity - Better protection for ecosystems, and more use of green infrastructure - More sustainable agriculture and forestry - Better management of fish stocks - Tighter controls on invasive alien species - A bigger EU contribution to averting global biodiversity loss

Methodology

Methodology for indicator calculation

SEBI ecosystem classes correspond to the following CORINE sub-classes:

(1)   Urban: 1.1.1., 1.1.2., 1.2.1., 1.2.2., 1.2.3., 1.2.4., 1.3.1., 1.3.2., 1.3.3., 1.4.1., 1.4.2.,

(2)   Cropland: 2.1.1., 2.1.2., 2.1.3., 2.2.1., 2.2.2., 2.2.3., 2.4.1.,

(3)   Grassland: 2.3.1., 3.2.1.,

(4)   Agricultural mosaics: 2.4.2, 2.4.3., 2.4.4.,

(5)   Woodland and forest: 3.1.1., 3.1.2., 3.1.3.,

(6)   Heathland and shrub: 3.2.2., 3.2.3.,

(7)   Transitional woodland: 3.2.4.,

(8)   Coastal: 3.3.1., 4.2.1., 4.2.2., 4.2.3., 5.2.1., 5.2.2.,

(9)   Sparsely vegetated land: 3.3.2., 3.3.3., 3.3.4., 3.3.5.,

(10)Inland wetlands: 4.1.1., 4.1.2.,

(11) Rivers and lakes: 5.1.1., 5.1.2.,

(12) Marine: 5.2.3.,

Methodology for gap filling

N/A

Methodology references

• SEBI 2010 Indicator Screening - Historical Development of Forest Area Based on UN Forest Resource Assessments. Van Brusselen, J. and Larsson T.B., 2005.  SEBI 2010 Expert Group 2, EEA, 2005
• European forest types - Categories and types for sustainable forest management reporting and policy. EEA, 2006.  EEA Technical report No 9/2006. European Environment Agency, Copenhagen, Denmark.
• Newly created MethodologyReference

Uncertainties

Methodology uncertainty

Datasets has been processed according to the land accounting methodology. Both for facilitating computation and visualising spatial change, land accounts are processed using a grid of 1x1 km. Each cell contains the exact CLC values but spatial aggregations are made of entire grid-cells, which may lead to some very limited marginal uncertainty for the border of a given national or regional land unit.

Differences in CLC change mapping technology (1990-2000 and 2000-2006):

In CLC1990-2000 changes were mapped by countries usually by intersecting CLC1990 and CLC2000 stock layers. The results were not always cleaned and non-changed parts might have remained in CLC 1990-2000 changes dataset. On the other hand, isolated changes below 25 ha could not be mapped by this technology. In CLC2000-2006 changes were mapped directly. This way all changes exceeding 5 ha were mapped and non-changed areas were better excluded from CLC-Changes.

Data sets uncertainty

Geographical and time coverage on EU level

Surfaces monitored with Corine Land Cover relate to the extension of urban systems that may include parcels not covered by construction, streets or other sealed surfaces. This is particularly the case for discontinuous urban fabric and recreation areas, which are considered as a whole. Monitoring the indicator with satellite images leads to the exclusion of small urban features in the countryside and most of the linear transport infrastructures, which are too narrow to be observed directly. Therefore, differences exist between CLC results and other statistics collected with different methodologies such as point or area sampling or farm surveys; this is often the case for agriculture and forest statistics. However, the trends are generally similar. The gap will be filled in at a further stage on the basis of a new high resolution database of transport infrastructures and calculations based on established coefficients for each type of transport.

Geographical and time coverage at the EU level:
All the EU-27 member states (except Greece) are covered with both CLC 2000 and 2006 results. Land cover changes in Liechtenstein remained below the detection level of Corine Land Cover change methodology. In most countries number of years between two CLCs is 6 years (with exception of Albania, Bosnia and Herzegovina, the Former Yugoslav Republic of Macedonia and Spain):
Albania (1995-2006) 11
Austria 6
Belgium 6
Bosnia and Herzegovina (1998-2006) 8
Bulgaria 6
Croatia 6
Cyprus 6
Czech Republic 6
Denmark 6
Estonia 6
Finland 6
Former Yugoslav Republic of Macedonia (1996-2006) 10
France 6
Germany 6
Hungary 6
Iceland 6
Ireland 6
Italy 6
Kosovo under UNSCR 1244/99 6
Latvia 6
Liechtenstein 6
Lithuania 6
Luxembourg 6
Malta 6
Montenegro 6
Netherlands 6
Norway 6
Poland 6
Portugal 6
Romania 6
Serbia 6
Slovakia 6
Slovenia 6
Spain (2000-2005) 5
Sweden 6
Switzerland 6
Turkey 6
United Kingdom 6



Representativeness of data on national level

At the national level, time differences between regions may happen in most countries and these are documented in the CLC meta data.

Rationale uncertainty

MAIN DISADVANTAGES OF THE INDICATOR

The use of remote sensing data implies that some degree of detail is lost. The Corine land cover data set is based on a minimal unit of 25 hectares and this implies that smaller areas of certain habitat types and linear features may not be adequately detected. Other data sets (e.g. statistical offices reporting for forests, cropland, grassland area) cannot be combined in this indicator calculation because the different definitions used as well as the different frequencies in updating will produce incomparable trends.