Keywords

1 HILDA

Product

LULC general

Dates

1900–2010 (every 10 years)

Formats

Raster

Pixel size

1 km

Thematic resolution

5 classes plus water:

1 (a), 1 (ag), 2 (v), 0 (m), 0 (na)Footnote 1

Compatible legends

IPCC, LCCS

Extent

European Union plus the UK and Switzerland

Updating

Not planned

Change detection

Yes

Overall accuracy

Not specified

Website of reference

Website Language English

https://www.wur.nl/en/Research-Results/Chair-groups/Environmental-Sciences/Laboratory-of-Geo-information-Science-and-Remote-Sensing/Models/Hilda.htm

Download site

https://www.wur.nl/en/Research-Results/Chair-groups/Environmental-Sciences/Laboratory-of-Geo-information-Science-and-Remote-Sensing/Models/Hilda/HILDA-data-downloads.htm

Availability

Format(s)

Open Access, after providing personal data

ESRI Grid, .tiff, .ascii

Technical documentation

Fuchs et al. (2013,2015a, b)

Other references of interest

Fuchs (2015)

Project

HIstoric Land Dynamics Assessment (HILDA) is a project aimed at reconstructing historic land cover/use and LUC changes in Europe. Unlike other LUC reconstruction projects and datasets, it allows us to study LUC changes over time. The recently launched HILDA + project takes the original project one step further by mapping historical LUC changes at a global scale for the period 1960–2019.

The reconstruction of historic LUC landscapes and changes is carried out using a model maintained and developed by the Department of Geoinformation Science and Remote Sensing of Wageningen University. The model allocates non-spatial historic LUC information on the ground.

Production method

Historic LUC maps for the HILDA project were obtained through an extensive workflow involving various steps. First, gross and net LUC changes per decade were obtained for the period 1950–2010 from a set of sources providing historic LUC information: UNFCCC national reporting data, CORINE Land Cover, Historisch Grondgebruik Nederland (HGN) for the Netherlands, FAO-RSS data and BioPress data with classified aerial photographs of 73 sample sites across Europe. Later, LUC data was spatially allocated by the HILDA model. Four categories were spatially allocated at this stage. A fifth category (other land) remained static throughout the time series. Water was a subclass of the “other land” category, which was only separated in the final maps for visualization purposes.

The model allocates the LUC categories using a series of probability maps. A specific probability map for each category was created on the basis of historical LUC maps and a range of socioeconomic and physical (soil properties, climate and terrain) factors. The categories were allocated hierarchically according to their socioeconomic value: settlements were allocated first, followed by croplands, forest and grasslands.

Once the model had been run for the 1950–2010 timeframe, four extra maps were obtained for the period 1900–1950 based on historical LUC statistics and an extrapolation of the change matrix. The pre-1950 maps therefore assume stable transition rates for the period 1950–2010. This could be an important source of uncertainty in these maps.

Product description

The product is delivered in four different packages, two of which include the series of LUC maps (1900–2010). Of these, one considers the net changes over the course of each decade, while the other considers the gross changes. The other two packages detail the specific transitions that take place between the different categories, one charting net changes and the other gross changes.

Each package can be downloaded in three different file formats (ESRI Grid, TIFF, ASCII). Each download includes a raster with LUC information for each decade and a supplementary file with the technical description of the product.

Downloads

Gross land changes

– Raster files with LUC maps for each decade

– Text document with technical information and the legend

Net land changes

– Raster files with LUC maps for each decade

– Text document with technical information and the legend

Transitions maps (for gross and net)

– Raster files with LUC maps for each decade

– Text document with technical information and the legend

Legend and codification

HILDA gross and net maps

Code

Label

Code

Label

111

Settlement

444

Grassland

222

Cropland

555

Other land

333

Forest

666

Water

HILDA gross and net transitions maps (1900–2000)

Code

Label

Code

Label

112

Cropland to settlement

242

Cropland to grassland

113

Forest to settlement

252

Cropland to other land

114

Grassland to settlement

262

Cropland to water

115

Other land to settlement

334

Grassland to forest

116

Water to settlement

335

Other land to forest

121

Settlement to cropland

336

Water to forest

131

Settlement to forest

343

Forest to grassland

141

Settlement to grassland

353

Forest to other land

151

Settlement to other land

363

Forest to water

161

Settlement to water

445

Other land to grassland

223

Forest to cropland

446

Water to grassland

224

Grassland to cropland

454

Grassland to other land

225

Other land to cropland

464

Grassland to water

226

Water to cropland

556

Water to other land

232

Cropland to forest

565

Other land to Water

HILDA gross and net transitions maps (2000–2010)

Code

Label

Code

Label

112

Settlement to cropland

242

Grassland to cropland

113

Settlement to forest

252

Other land to cropland

114

Settlement to grassland

262

Water to cropland

115

Settlement to other land

334

Forest to grassland

116

Settlement to water

335

Forest to other land

121

Cropland to settlement

336

Forest to water

131

Forest to settlement

343

Grassland to forest

141

Grassland to settlement

353

Other land to forest

151

Other land to settlement

363

Water to forest

161

Water to settlement

445

Grassland to other land

223

Cropland to forest

446

Grassland to water

224

Cropland to grassland

454

Other land to grassland

225

Cropland to other land

464

Water to grassland

226

Cropland to water

556

Other land to Water

232

Forest to cropland

565

Water to other land

Practical considerations

This is a valuable dataset because of the rich historic LUC information it provides. There are very few long, dense historical series of LUC maps that measure LUC change over time. Nonetheless, users should be aware of the uncertainties associated with this dataset. The maps prior to 1950 were created by extrapolating the patterns of change for the period 1950–2010. This could introduce a high degree of uncertainty.

An online visualization of the maps for the years 1900 and 2010 is available, together with other auxiliary information, at http://www.geo-informatie.nl/fuchs003/.

To study global historical LUC change at a similar level of detail, users should refer to the associated HILDA+ project.

2 CLC—CORINE Land Cover

Product

LULC general

Dates

1990, 2000, 2006, 2012, 2018

Formats

Vector and raster

Scale/Pixel size

Photointerpretation scale: 1:100,000

Minimum Mapping Unit: 25 ha/5 ha for changes

Minimum Mapping Width: 100 m

Pixel size (raster): 100 m

Thematic resolution

44 classes:

11 (a), 8 (ag), 8 (v), 6 (m), 3 (na)

Compatible legends

CLC

Extent

Europe, with an increasing number of countries taking part in the project each year (39 in CLC18)

Updating

Scheduled updates every 6 years

Change detection

Yes, through the layer of changes

Overall accuracy

Expected to be >85%

Website of reference

Website Language English, German and French

https://land.copernicus.eu/pan-european/corine-land-cover

Download site

https://land.copernicus.eu/pan-european/corine-land-cover

Availability

Format(s)

Open Access previous registration

.tiff, .gdb, .gpkg

Technical documentation

Bossard et al. (2000), Büttner et al. (2002, 2011, 2012, 2014) European Environment Agency (1994, 2006a, b, 2007), Jaffrain et al (2017), Kosztra et al. (2019), Soukup et al. (2017)

Other references of interest

Bach et al. (2006), Bielecka and Jenerowicz (2019), Büttner (2014), European Environment Agency (2006c), Feranec et al. (2010, 2016), Gallego (2001), García-Álvarez and Camacho Olmedo (2017), Neumann et al. (2007)

Project

CORINE Land Cover (CLC) is a European project monitoring Land Use and Cover that dates back to 1985. It aims to map land uses and land covers across the whole continent according to the same rules. It is currently part of the land monitoring efforts of the Copernicus programme.

The number of countries taking part in the project has been increasing since its inception, from the initial group of 26 countries that created the CLC 1990 to the 39 countries that participated in the most recent editionFootnote 2. In the meantime, the production of CLC has undergone several technical and methodological changes. The fact that CLC is produced at a national level means that methods vary from one country to the next.

Because of its long life, detail, consistency and wide range of applications, CLC is one of the most renowned LUC mapping initiatives worldwide. Various European countries have developed national LUC products based on CLC. In some cases, these products are new CLC layers with an extended legend, adapted to the specificities of the country. In other cases, they are new CLC layers for different dates to those used in the main Europe-wide project.

Production method

The production of CLC is coordinated by the European Environment Agency (EEA). Each participant country is responsible for mapping its own territory according to the general guidelines developed by the EEA.

The method of production may vary from country to country. Initially, CLC was mapped at national scales based on the photointerpretation of Landsat imagery. In the following editions, most of the countries decided to stick to this method, using different satellite imagery according to EEA prescriptions: Landsat, SPOT; ITS P6, RapidEye, LISS III, Sentinel. In the latest editions, the production method has varied in some cases. A few countries, like Germany or Spain, produce the CLC database by generalizing national LUC databases at finer scales. This has introduced important changes in the way land uses and covers are mapped over time for these countries. For both production methods, photointerpretation and map generalization, the CLC map obtained is then subject to expert review to ensure its consistency and validity.

The first CLC map was produced for the reference year 1990 and the subsequent editions have been updates of this initial map. The national teams do not draw a new map for each new reference year. Instead, they map the changes for the analysed period (e.g. 1990–2000) and then update the base map for the new reference year. In this updating process, any errors detected in the base map are also corrected. If important changes have been made in the CLC production method, the base map is also updated according to the new method.

In addition to the maps for each reference year, CLC produces change layers for each period between reference years: 1990–2000, 2000–2006, 2006–2012, 2012–2018. The maps showing changes do not follow the same mapping rules as the base CLC maps and show more information than the base layers for the reference years (MMU of 5ha). The CLC production team therefore recommends that LUC changes be studied using these change layers, rather than by cross-tabulating and comparing base CLC maps.

Product description

CLC is made up of two spatial layers: a Land Use Cover map for each reference year (1990, 2000, 2006, 2012, 2018) and a layer of Land Use Cover changes for each analysis period (1990–2000, 2000–2006, 2006–2012, 2012–2018). The reference map for each year provides Land Use Cover information for the total area of the participant countries. The map of changes only accounts for the changes that took place in the period under consideration. Rather than comparing two reference maps, the CLC layer of changes maps all changes bigger than 5ha and discards all technical changes that did not take place on the ground.

CLC layers are provided in either vector (ESRI or GeoPackage databases) or raster (.tiff) formats. As might be expected, the vector data is much heavier than the raster data, because of its higher definition.

Together with the LUC layers, the CLC product includes all the auxiliary information required to understand the LUC information provided by the CLC layers: a style layer for the raster, the legend description, technical information and other relevant metadata. LUC maps for the French overseas departments (Guadeloupe, French Guinea, Martinique, Mayotte and Reunion) are also provided in auxiliary layers.

Downloads

The base layers with LUC maps for each reference year (CLC) have the same structure and group of files, as do the change layers for each period of analysis (CHA). This is why we only describe the file structure once for each type of format.

CLC 2018 (Geodatabase)/CHA 2012–2018 (Geodatabase)

– Geodatabase files with CLC vector layers (DATA folder)

– Folder with CLC vector data for French overseas departments

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Legend folder)

– Excel presenting the CLC legend, including information about the RGB colours for each class (Legend folder)

– Text documents describing the CLC legend, including information about the RGB colours for each class (Legend folder)

– Folder with metadata files (.xml)

– PDF and Excel sheet with information about CLC country coverage (Documents folder)

– A Word document explaining how to use the CLC files for the product in QGIS (Documents folder)

– Three text documents with technical information about the CLC layers (Documents folder)

CLC 2018 (GeoPackage)/CHA 2012–2018 (GeoPackage)

– GeoPackage file with CLC vector layers (DATA folder)

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Legend folder)

– Excel presenting the CLC legend, including information about the RGB colours for each class (Legend folder)

– Text documents describing the CLC legend, including information about the RGB colours for each class (Legend folder)

– Folder with metadata files (.xml)

– PDF and Excel sheet with information about CLC country coverage (Documents folder)

– A Word document explaining how to use the CLC files for the product in QGIS (Documents folder)

– PDFs and text documents with technical information about the CLC layers (Documents folder)

CLC 2018 (Raster)/CHA 2012–2018 (Raster)

– Raster file with CLC map (DATA folder)

– Folder with CLC raster data for French Overseas Departments (DATA folder)

– Layer style files for ArcGIS (.lyr) and QGIS (.qml) (Legend folder)

– Layer style files for ArcGIS (.lyr) and QGIS (.qml) for French Overseas Departments (French_DOMs folder)

– Text document describing the CLC legend, including information about the RGB colours for each class (Legend folder)

– Folder with metadata files (.xml)

– PDF and Excel sheet with information about CLC country coverage (Documents folder)

– A Word document explaining how to use the CLC files for the product in QGIS (Documents folder)

– PDFs and text documents with technical information about the CLC layers (Documents folder)

Database

CLC 2018

– OBJECTID: Unique identifier for each polygon.

– Code_18: LUC code for the year 2018.

– Remark

– Area_Ha: Area of the polygon, in hectares.

– ID: Unique identifier for each polygon.

– Shape_Length: Perimeter of the polygon, in metres.

– Shape_Area: Area of the polygon, in square metres.

– C18: LUC code for the year 2018.

CHA 2012–2018

– OBJECTID: Unique identifier for each polygon

– Change: Change code made up of the CLC code for the oldest year (on the right) and the CLC code for the most recent year on the left (2018)

– ID:

– Code_12: LUC code for the year 2012

– Code_18: LUC code for the year 2018

– Chtype

– Remark

– AREA_HA: Area of the polygon, in hectares

– Shape_Length: Perimeter of the polygon, in metres

– Shape_Area: Area of the polygon, in square metres

Legend and codification

Code

Label

Code

Label

111

Continuous urban fabric

313

Mixed forest

112

Discontinuous urban fabric

321

Natural grasslands

121

Industrial or commercial units

322

Moors and heathland

122

Road and rail networks and associated land

323

Sclerophyllous vegetation

123

Port areas

324

Transitional woodland-shrub

124

Airports

331

Beaches, dunes, sands

131

Mineral extraction sites

332

Bare rocks

132

Dump sites

333

Sparsely vegetated areas

133

Construction sites

334

Burnt areas

141

Green urban areas

335

Glaciers and perpetual snow

142

Sport and leisure facilities

411

Inland marshes

211

Non-irrigated land

412

Peat bogs

213

Rice fields

421

Salt marshes

221

Vineyards

422

Salines

222

Fruit trees and berry plantations

423

Intertidal flats

223

Olive groves

511

Water courses

231

Pastures

512

Water bodies

241

Annual crops associated with permanent crops

521

Coastal lagoons

242

Complex cultivation patterns

522

Estuaries

243

Land principally occupied by agriculture, with significant areas of natural vegetation

523

Sea and ocean

244

Agro-forestry areas

999

NO DATA

311

Broad-leaved forest

990

UNCLASSIFIED LAND SURFACE

312

Coniferous forest

995

UNCLASSIFIED WATER BODIES

Practical considerations

CLC was originally mapped in vector format. This format provides higher precision and detail and is therefore recommended when working at local and regional scales. At national and supranational scales, raster data can be more suitable, as vector data is too heavy and may be difficult to handle in desktop computers with insufficient processing power.

Users can download the vector CLC to rasterize the database to the spatial resolution they require. The 100 m offered is the reference resolution provided by the EEA, but it is not the only one at which the map could be used.

Users should be aware that different mapping methodologies were used in different countries, and in some countries, at different times. This could result in significant differences in the way the landscape is mapped and conceptualised, which could introduce important sources of uncertainty in our studies and analyses. The same category could be interpreted differently in different countries, and even within the same country, a particular category could be mapped differently at different times if the production method changes. Those wishing to analyse LUC change should therefore use the change layers rather than the maps.

3 PELCOM—Pan-European Land Use and Land Cover Monitoring

Product

LULC general

Dates

1997

Formats

Raster

Pixel size

1 km

Thematic resolution

16 classes:

1 (a), 3 (ag), 5 (v), 1 (m), 2 (na)

Compatible legends

No

Extent

Europe

Updating

No

Change detection

No (only one date)

Overall accuracy

Expected to be >69%

Website of reference

Website Language English

http://www.geo-informatie.nl/projects/pelcom/

Download site

http://www.geo-informatie.nl/projects/pelcom/public/index.htm

Availability

Format(s)

Open Access

.tiff

Technical documentation

Champeaux et al. (2000), Mücher (2000), Mücher et al. (2000)

Other references of interest

Project

PELCOM (Pan-European Land Cover Monitoring) was a research project funded by the European Union that ran from 1996 to 1999. The main purpose of the project was to develop a consistent methodology to create a continental LUC map for Europe from remote sensing sources. Users were consulted about their needs and requirements and revealed that they would like to have LUC data at coarser and finer spatial resolutions than CLC, and that CLC could be updated more frequently. They also made clear that a dataset of this kind would be useful for environmental modelling and monitoring purposes.

At the time the project was launched, no consistent continental LUC maps were available at high spatial resolution (at least 1 km). The map created through the project sought to provide a high-resolution continental LUC dataset that could later be updated frequently. However, despite these original intentions, the PELCOM map has not been updated since the project came to an end.

Production method

The classification carried out for the PELCOM map was based on AVHRR imagery and NDVI composites from the DLR archive of the JRC. An improved stratified, integrated classification methodology was specifically developed by the creators of this map. To this end, Europe was divided into different strata according to similarities in LULC patterns and phenology.

The classification process consisted of several steps, in which users played an important role. Both supervised and unsupervised classifiers were employed. Some classes (forest, water bodies, urban areas) were mapped through specific workflows, using masks and other strategies, to improve the uncertainty and errors associated with their classification.

Product description

PELCOM may be downloaded in three different formats: ESRI-grid, ERDAS-Image and ENVI. The download includes the raster with the LUC map and, depending on the format chosen, auxiliary information about the product (readme and symbology files).

Detailed technical documentation about the map and its production method is also available from the download site.

Downloads

PELCOM ESRI-grid

– Raster file with LUC map

– Preview image of the product

– Readme file with information about the product (.doc)

– File with raster symbology for ArcGIS (.avl)

Legend and codification

Code

Label

Code

Label

11

Coniferous forest

60

Barren land

12

Deciduous forest

70

Permanent Ice & Snow

13

Mixed forest

80

Wetlands

20

Grassland

91

Inland waters

31

Rainfed arable land

92

Sea

32

Irrigated arable land

100

Urban areas

40

Permanent crops

110

Data gaps

50

Shrubland

111

Out of scope

4 Annual Land Cover Product

Product

LULC general

Dates

2000–2019

Formats

Raster

Pixel size

30 m

Thematic resolution

33 classes:

8 (a), 7 (ag), 7 (v), 1 (m), 0 (na)

Compatible legends

LUCAS, CLC

Extent

Europe

Updating

Not planned

Change detection

Not recommended

Overall accuracy

Evaluation in process

Website of reference

Website Language English

https://medium.com/swlh/europe-from-above-space-time-machine-learning-reveals-our-changing-environment-1b05cb7be520

Download site

https://maps.opendatascience.eu/

Availability

Format(s)

Open Access

.tiff

Technical documentation

Not published yet

Other references of interest

Project

An open annual land cover dataset for Europe has been produced in the context of the “Geo-harmonizer: EU-wide automated mapping system for harmonization of Open Data based on FOSS4G and Machine Learning”, a project coordinated by the Czech Technical University in Prague. This project is part of the Connecting Europe Facility (CEF) in Telecom, which aims to deploy digital service infrastructures (DSIs) that can facilitate cross-border interaction between public administrations, businesses and citizens.

The Geo-harmonizer project has developed a web-based system (Open Data Science Europe) that hosts open European thematic geospatial layers, including one on land cover. They were specifically created for the project from other data sources for the period 2000–2020 using modelling techniques. These harmonized European layers overcome the limitations resulting from the use of national datasets that were created with different parameters and have different characteristics.

Apart from a layer on land cover, Open Data Science Europe hosts data on subjects such as the environment, terrain, clime, soils or vegetation. These data are complementary to the datasets provided by the Copernicus Land Monitoring Service, also at continental level.

The project has the same values and approach as other Open Science projects in the geospatial field, such as Open Land Map and Open Street Map.

Production method

Open Data Science Europe’s Annual Land Cover Product is obtained by producing a series of probability layers for each of the 33 LUC categories that were mapped. The land cover with the highest probability for each year and pixel according to these layers was the one finally selected to create the general LUC maps.

Probability layers were obtained through a set of three Machine Learning (ML) models: Random Forest, XGBoost and Artificial Neural Network. The models were trained with reference data obtained from CLC and LUCAS and input Landsat imagery (LANDSAT ARD), night lights data (VIIRS/SUOMI NPP), Global surface water frequency and an EU DTM.

A final probability layer for each LUC category was obtained after running a Logistic regression classifier on the results of three ML models. The uncertainty of the probability layers for each LUC category was also calculated as the standard deviation of the three predicted probabilities from the ML models.

Product description

The dataset can be individually downloaded for each available year of the period 2000–2019 from the Open Data Science Europe viewer. The download contains the raster file with the LUC information, but offers no other auxiliary data. Nonetheless, a layer style file to symbolize the dataset in QGISFootnote 3 can be downloaded separately.

Downloads

Annual Land Cover Product 2019

– Raster file with LUC map (.tiff)

Legend and codification

Code

Label

Code

Label

111

Urban fabric

321

Natural grasslands

122

Road and rail networks and associated land

322

Moors and heathland

123

Port areas

323

Sclerophyllous vegetation

124

Airports

324

Transitional woodland-shrub

131

Mineral extraction sites

331

Breaches, dunes, sands

132

Dump sites

332

Bare rocks

133

Construction sites

333

Sparsely vegetated areas

141

Green urban areas

334

Burnt areas

211

Non-irrigated arable land

335

Glaciers and perpetual snow

212

Permanently irrigated arable land

411

Inland wetlands

213

Rice fields

421

Maritime wetlands

221

Vineyards

511

Water courses

222

Fruit trees and berry plantations

512

Water bodies

223

Olive groves

521

Coastal lagoons

231

Pastures

522

Estuaries

311

Broad-leaved forest

523

Sea and ocean

312

Coniferous forest

  

Practical considerations

The dataset is currently only available for download at the Opendatascience website. However, it will soon be uploaded to public repositories, where users will be able to access all data from the project, including layers of uncertainty. Information about the dataset production procedure will also be published in the coming months together with other relevant information.

The dataset can also be accessed through a WFSFootnote 4 and a file service (Cloud-Optimized GeoTIFFs)Footnote 5 in QGIS or other common GIS software. The map producers also provide information about how to access the data through GDAL, R and Python. This can be found by clicking on the About tab in the Opendatascience website.

5 GlobCorine

Product

LULC general

Dates

2005, 2009

Formats

Raster

Pixel size

300 m

Thematic resolution

17 classes:

1 (a), 3 (ag), 7 (v), 4 (m), 1 (na)

Compatible legends

CLC—FAO LCCS

Extent

Continental (Europe and surroundings)

Updating

No

Change detection

Not recommended

Overall accuracy

>48% or >79% depending on the validation dataset considered

Website of reference

Website Language English

http://due.esrin.esa.int/page_project114.php

Download site

https://maps.elie.ucl.ac.be/geoportail/

Availability

Format(s)

Open Access (only 2005 map)

.tiff

Technical documentation

Bontemps et al. (2010), Defourny et al. (2010a b, c)

Other references of interest

Bontemps et al. (2009)

Project

Based on earlier efforts in GlobCover, the ESA launched the GlobCorine project in collaboration with the European Environment Agency (EEA) and the Université Catholique de Louvain (UCL). The aim was to create a new LUC product for the European continent that was compatible with the Corine Land Cover (CLC) classification and built on the work already carried out as part of the GlobCover project.

Production method

GlobCorine was produced by classifying the same MERIS imagery used for GlobCover. The same production method was used in the two LUC maps available and was similar to the one already used for GlobCover. It consisted of a series of supervised and unsupervised classification routines to identify spectro-temporal classes. These were later automatically labelled with the information provided by auxiliary datasets, mainly Corine Land Cover (CLC) and GlobCover. For classification purposes, the world was divided into different regions according to their ecological and reflectance characteristics.

An extra classification was carried out for mixed categories. The final LUC maps were then corrected and improved in a post-classification phase with the help of auxiliary data and expert knowledge.

Product description

Only one of the two GlobCorine maps is currently available for download: the map for the reference year 2005. The download includes the raster with the LUC map, the legend, a file to symbolize it in GIS software and all relevant technical information explaining the characteristics of the dataset.

Downloads

GlobCorine 2005

– Raster file with LUC map (GLOBCORINE_LC)

– Preview image of the product (GLOBCORINE_LC)

– Layer style files for ArcGIS (.lyr) and ENVI (.dsr) (GLOBCORINE_LC)

– Excel sheet with the map legend (“GlobCorine_legend”) (GLOBCORINE_LC)

– PDFs with technical information about the product (Documentation)

– PDF with a description of the downloaded product (README)

Legend and codification

Code

Label

Code

Label

10

Urban areas and associated areas

100

Complex cropland

20

Rainfed cropland

110

Mosaic cropland (50–80%) / natural vegetation (20–50%)

30

Irrigated cropland

120

Mosaic natural vegetation (50–80%) / cropland (20–50%)

40

Forest

130

Mosaic herbaceous (50–80%) / shrub-trees (20–50%)

50

Shrubland

140

Mosaic shrub-trees (50–80%) / herbaceous (20–50%)

60

Grassland

200

Water bodies

70

Sparsely vegetated areas

210

Permanent snow and ice

80

Vegetated low-lying areas on regularly flooded soil

220

No Data

90

Bare areas

  

Practical considerations

The product is no longer available for download from the official website of the EEA. The only edition that can still be obtained is the map for 2005, which is available through the geoportal of the Université Catholique de Louvain, one of the producers of the dataset. The map can also be consulted online at the same website, without having to download it.

While the GlobCorine classification legend focuses particularly on land use, GlobCover centres on land cover. GlobCorine can therefore be regarded as a complementary dataset to GlobCover.

6 Urban Atlas

Product

LULC general

Dates

2006, 2012, 2018

Formats

Vector

Scale

Photointerpretation scale: 1:10,000

Minimum Mapping Unit: 0.25ha in urban areas and 1ha in rural areas

0.1ha for urban changes and 0.25ha for rural/natural changes

Minimum Mapping Width: 10 m

Thematic resolution

29 classes:

17 (a), 4 (ag), 2 (v), 2 (m), 2 (na)

Compatible legends

CLC

Extent

Europe (39 countries)

Updating

Every 6 years

Change detection

Through map of changes

Overall accuracy

Expected to be > 80%

Website of reference

Website Language English, German and French

https://land.copernicus.eu/local/urban-atlas

Download site

https://land.copernicus.eu/local/urban-atlas

Availability

Format(s)

Open Access under registration

.gpkg

Technical documentation

Copernicus Programme (2020), Gallaun (2017), Hirschmugl et al. (2018), Silva et al. (2013, 2016)

Other references of interest

Barranco et al. (2014), European Commission and OECD (2012), Jaffrain et al. (2016), Montero et al. (2014), Petrişor and Petrişor (2015), Prastacos et al. (2011), Seifert (2009)

Project

Urban Atlas is part of the Copernicus programme and provides very detailed LUC information for Functional Urban Areas (FUA) in Europe. A Functional Urban Area (as defined by the European Commission and the OECD) is an urban space that joins the core areas of cities with their surrounding commuter belts.

The Urban Atlas aims to contribute to the study of urban areas and their dynamics, in line with the needs of the European Commission and other European initiatives, such as ESPON and INTERREG. It therefore has a clear goal to inform policy-making.

Three editions of the Urban Atlas have so far been published, with more FUAs participating in each one. 319 FUAs were mapped for reference year 2006, 785 for 2012 and 788 for 2018. New updates of the Urban Atlas are expected every 6 years.

For each edition, a detailed LUC map of the FUAs is provided, as well as a map of the changes that have taken place over the period under consideration (e.g. 2006–2012). A Street Tree Layer map is also provided for the 2012 and 2018 editions. The 2012 Urban Atlas includes a building height map for core areas (not FUA) of European capitals in the EEA39. Polygons of the 2012 Urban Atlas also include population estimates.

Production method

Urban Atlas is obtained through automatic classification and manual photointerpretation of high-resolution satellite imagery: the optical VHR coverage of the Copernicus programme, at a spatial resolution of 2–4 m.

First, the imagery is automatically segmented and classified, differentiating between basic land cover classes. Later, the detailed interpretation of land cover classes is carried out visually. A range of auxiliary data are applied in this process: topographic maps, the High Resolution Layer for impervious surfaces, road networks from COTS (Commercial Off-The-Shelf) navigation data and OSM as well as other data sources depending on the class under consideration (e.g. Google Earth, local city maps, cadastral data or very high resolution imagery, at a spatial resolution of up to 1 m).

Change detection for each period is carried out independently, based on the Urban Atlas map for the previous year and a combination of both automatic and manual approaches for change detection. In the change detection process, misclassifications for the previous year of reference are corrected.

There are certain exceptions to the Minimum Mapping Units and Minimum Mapping Widths, depending on the characteristics and pattern of the class being analysed. However, no features are mapped below the 0.5ha threshold.

Product description

Urban Atlas is distributed in single files for each FUA. There is no single common file that hosts all the FUAs together. A different file must be downloaded for each year and for each available change layer.

Downloads include the vector layers in Geopackage format with the LUC information, the boundaries of the FUAs and their urban cores, a metadata file and layer style files to symbolize the vector layers in GIS.

For reference years 2012 and 2018, the Street Tree Layer can be downloaded for each FUA. This layer represents contiguous rows or patches of trees covering at least 0.5ha. For the reference year 2012, a building height model in raster format can also be downloaded.

Downloads

Urban Atlas 2018 (Madrid)

– GeoPackage file with Urban Atlas vector layers: Urban Atlas 2018, Urban Core and Boundary (DATA folder)

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Legend folder)

– Metadata file (.xml) (Metadata folder)

Urban Atlas changes 2012–2018 (Madrid)

– GeoPackage file with Urban Atlas Change vector layers: Urban Atlas Change 2012–2018, Urban Core and Boundary (DATA folder)

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Legend folder)

– Metadata file (.xml) (Metadata folder)

Street tree layer 2012 – STL (Madrid)

– Vector file with STL layer

– Vector file with FUA boundary

Building height 2012 (Madrid)

– Raster file with building heights (DATA folder)

– PDFs with technical information about the product (DOC folder)

– Metadata file (.xml) (Metadata folder)

Database

Urban area 2018 (Madrid)

– FID: Unique identifier for each polygon

– Country: Country code

– FUA_name: Name of the Functional Urban Area

– FUA_code: Code for the Functional Urban Area

– Code_2018: LUC code for the year 2018

– Class_2018: LUC description for the year 2018

– Prod_date: Map production year

– Identifier: Unique identifier for each polygon

– Perimeter: Perimeter of the polygon, in metres

– Area: Area of the polygon, in square metres

– Comment: Extra field for additional comments about the mapped features

Urban area change 2012–2018 (Madrid)

– FID: Unique identifier for each polygon

– Country: A two-letter code to identify each country

– FUA_name: Name of the Functional Urban Area

– FUA_code: Code for the Functional Urban Area

– Code_2018: LUC code for the year 2018

– Class_2018: LUC description for the year 2012

– Prod_date: Map production year

– Identifier: Unique identifier for each polygon

– Perimeter: Perimeter of the polygon, in metres

– Area: Area of the polygon, in square metres

– Comment: Extra field for additional comments about the mapped features

– Code_2012: LUC code for the year 2012

– Class_2012: LUC description for the year 2012

Street tree layer 2012—STL (Madrid)

– COUNTRY: A two-letter code for each different country

– CITIES: Name of the Functional Urban Area

– FUA_OR_CIT: Code of the Functional Urban Area

– STL: Street Tree Layer code

– Shape_Leng: Perimeter of the polygon, in metres

– Shape_Area: Area of the polygon, in square metres

Legend and codification

Urban Atlas

Code

Label

Code

Label

11100

Continuous urban fabric (S.L. > 80%)

14100

Green urban areas

11210

Discontinuous dense urban fabric (S.L. 50–80%)

14200

Sports and leisure facilities

11220

Discontinuous medium-density urban fabric (S.L. 30–50%)

21000

Arable land (annual crops)

11230

Discontinuous low-density urban fabric (S.L. 10–30%)

22000

Permanent crops

11240

Discontinuous very low-density urban fabric (S.L. < 10%)

23000

Pastures

11300

Isolated structures

24000

Complex and mixed cultivation

12100

Industrial, commercial, public, military and private units

25000

Orchards

12210

Fast transit roads and associated land

31000

Forests

12220

Other roads and associated land

32000

Herbaceous vegetation associations

12230

Railways and associated land

33000

Open spaces with little or no vegetation

12300

Port areas

40000

Wetlands

12400

Airports

50000

Water

13100

Mineral extraction and dump sites

91000

No data (Clouds and shadows)

13300

Construction sites

92000

No data (Missing imagery)

13400

Land without current use

  

Street tree layer

Code

Land cover

1

Tree cover

Practical considerations

LUC change must be analysed using the change layer. Comparing Urban Atlases for different years of reference will highlight many technical changes that did not actually happen on the ground.

The Urban Atlas product can be also consulted online at the download webpage.

7 N2K—Natura 2000

Product

LULC general

Dates

2006, 2012, 2018

Formats

Vector

Scale

Photointerpretation scale: 1:5,000–1:10,000

Minimum Mapping Unit: 0.5 ha

Minimum Mapping Width: 10 m

Thematic resolution

48 classes:

8 (a), 6 (ag), 13 (v), 9 (m), 0 (na)

Compatible legends

Urban Atlas, Riparian Zones, Coastal Zone product

Extent

Europe (29 countries)

Updating

Not planned

Change detection

Yes

Overall accuracy

Expected to be >80%

Website of reference

Website Language English, German and French

http://land.copernicus.eu/local/natura

Download site

http://land.copernicus.eu/local/natura

Availability

Format(s)

Open Access under registration

.gdb, .gpkg

Technical documentation

Buck and Büscher (2018)

Other references of interest

Project

N2K was developed as part of the Copernicus Land Monitoring programme. It maps land uses and covers in the areas that form part of the Natura 2000 network, plus a 2 km buffer zone around their perimeters. Natura 2000 is a network that protects natural areas with rare and threatened species or with rare types of natural habitat.

The dataset first appeared in 2015. A reviewed edition was issued in 2017 with a new classification legend that made it compatible with other European local reference LUC datasets: Riparian Zones, N2K and the Coastal Zone product.

Production method

N2K is obtained by photointerpretation of high-resolution imagery. Various auxiliary datasets are used in the photointerpretation process, namely CORINE Land Cover, Urban Atlas, High Resolution Layers, topographic maps, national WMS services and COTS navigation data. The changes are also photointerpreted by comparing satellite images at two different points in time.

Product description

N2K is distributed as a single vector file covering all mapped Nature 2000 areas. Two formats are available: ESRI Geodatabase and Geopackage. Downloads include the layers with LUC information, a style file to symbolize the layers in GIS and a pdf with the product classification scheme.

Downloads

N2K 2012 (Geodatabase)

– Geodatabase files with N2K vector layers

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Legend folder)

– Metadata file (.xml) (Metadata folder)

– PDF with nomenclature guidelines

N2K 2012 (GeoPackage)

– GeoPackage files with N2K vector layers

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Legend folder)

– Metadata file (.xml) (Metadata folder)

– PDF with nomenclature guidelines

Database

Product: N2K 2000

– OBJECTID: Unique identifier for each polygon

– ID

– UID

– SITECODE

– GRASSTYPE

– MAES_1_12: MAES class Level 1 for 2012

– MAES_2_12: MAES class Level 2 for 2012

– MAES_3_12: MAES class Level 3 for 2012

– MAES_4_12: MAES class Level 4 for 2012

– COMMENT_12: Comments on the 2012 mapping

– NODATA_12: Objects with no data in 2012

– MAES_1_06: MAES class Level 1 for 2006

– MAES_2_06: MAES class Level 2 for 2006

– MAES_3_06: MAES class Level 3 for 2006

– MAES_4_06: MAES class Level 4 for 2006

– COMMENT_06: Comments on the 2006 mapping

– NODATA_06: Objects with no data in 2006

– CHANGECODE: 2006–2012 changes

– AREA_HA: Area of the polygon, in hectares

– ID: unique identifier for each polygon

Legend and codification

N2K was produced according to a hierarchical classification legend made up of four different levels, the most detailed of which is provided here (MAES L3). Information about the other levels of classification and their codes can be found in the technical documents accompanying the dataset.

Code

Label

Code

Label

111

Urban fabric (predominantly public and private units)

4211

Semi-natural grassland with woody plants (C.C.D. ≥ 30%)

112

Industrial, commercial and military units

4212

Semi-natural grassland without woody plants (C.C.D. ≤ 30%)

121

Road networks and associated land

422

Alpine and sub-alpine natural grassland

122

Railways and associated land

511

Heathland and Moorland

123

Port areas and associated land

512

Other scrub land

124

Airports and associated land

621

Beaches and dunes

131

Mineral extraction, dump and construction sites

622

River banks

132

Land without current use

631

Bare rocks and rock debris

211

Arable land

632

Burn areas (except burnt forest)

212

Greenhouses

633

Glaciers and perpetual snow

221

Vineyards, fruit trees and berry plantations

721

Exploited peat bog

222

Olive groves

722

Unexploited peat bog

231

Annual crops associated with permanent crops

811

Coastal salt marshes

232

Complex cultivation patterns

812

Salines

233

Land principally occupied by agriculture with significant areas of natural vegetation

813

Intertidal flats

234

Agro-forestry

821

Coastal lagoons

311

Natural and semi-natural broadleaved forest

822

Estuaries

312

Highly artificial broadleaved plantations

911

Interconnected water courses

321

Natural and semi-natural coniferous forest

912

Highly modified water courses and canals

322

Highly artificial coniferous plantations

913

Separated water bodies belonging to the river system

331

Natural and semi-natural mixed forest

921

Natural water bodies

332

Highly artificial mixed plantations

922

Artificial standing water bodies

341

Transitional woodland and scrub

923

Intensively managed fish ponds

342

Lines of threes and scrubs

924

Standing water bodies of extractive industrial sites

Practical considerations

N2K files are very heavy (over 2gb), which means that they may be difficult to use for those without powerful computers. The map can also be consulted online in a viewer included in the download website of the product.

8 Riparian Zones Land Cover/Land Use—Riparian Zones (RZ)

Product

LULC general

Dates

2012, 2018

Formats

Vector

Scale

Photointerpretation scale: 1:10,000

Minimum Mapping Unit: 0.5 ha

Minimum Mapping Width: 10 m

Thematic resolution

56 classes:

11 (a), 6 (ag), 16 (v), 10 (m), 0 (na)

Compatible legends

Urban Atlas, N2K, Coastal Zone product

Extent

Europe (39 countries)

Updating

Not expected

Change detection

No (only one date)

Overall accuracy

Expected to be >85%

Website of reference

Website Language English, French and German

https://land.copernicus.eu/local/riparian-zones

Download site

https://land.copernicus.eu/local/riparian-zones/land-cover-land-use-lclu-image?tab=download

Availability

Format(s)

Open Access after registration

.shp

Technical documentation

Tamame et al. (2018), Vandeputte et al. (2018), Weissteiner et al. (2016)

Other references of interest

Piedelobo et al. (2019), Ugille (2019)

Project

Riparian Zones (RZ) is one of the local datasets produced as part of the Copernicus Land Monitoring Programme. It focuses on riparian areas (i.e. transitional areas between land and freshwater ecosystems with very specific characteristics) associated with Strahler level 2–8 rivers.

This product was created to support the Mapping and Assessment of Ecosystems and their Services (MAES) within the context of the EU Biodiversity Strategy for 2020. It is also intended for use in relation to the Habitats, Birds and Water Framework Directives.

The Riparian Zones dataset was initially launched in 2015, with an extension in 2017/18 to include riparian areas from Strahler 2 rivers. In 2017 the classification scheme was adapted to make it compatible with other local products developed under the Copernicus Land Monitoring framework.

Together with the LUC map of riparian zones, two extra complementary products are also provided: a delineation of Riparian Zones based on a fuzzy modelling approach and an inventory of the Green Linear Elements (hedgerows and lines of trees) growing in those riparian areas.

Production method

The RZ map was obtained through semi-automatic classification of very high-resolution imagery captured by the SPOT and Pleiades satellites (1.5–2.5 m). This classification was later refined with the aid of visual interpretation and intersected with the following auxiliary datasets: CORINE Land Cover, Imperviousness HRL, Tree Cover Density HRL and Urban Atlas.

Product description

A different vector file is provided for each riparian area mapped. Downloads include the vector file with LUC information and pdf documents with information about the product.

Downloads

Riparian zones land cover land use 2012 (vector)

– Vector files with LUC information (Data folder)

– PDF with nomenclature guidelines (Documents folder)

– PDF with product specifications (Documents folder)

– Metadata files (.xml) (Metadata folder)

Database

Riparian zones land cover land use 2012—LCLU (vector)

– ID: Unique identifier for each polygon

– DU_ID: Mapped area Code

– MAES_1: MAES class Level 1

– MAES_2: MAES class Level 2

– MAES_3: MAES class Level 3

– MAES_4: MAES class Level 4

– UA

– AREA_HA: Area of the polygon, in hectares

– NODATA: Unclassifiable areas due to clouds, shadows, snow, haze or missing data

– COMMENT: Comment field for additional information

Legend and codification

The Riparian Zones dataset was produced following a hierarchical classification legend made up of three different levels, the most detailed of which is provided here. Information about the other levels of classification of LUC categories can be found in the technical documents accompanying the dataset.

Code

Label

Code

Label

1111

Continuous Urban Fabric (IM.D ≥ 80%)

41

Managed grassland

1112

Dense Urban Fabric (IM.D ≥ 30–80%)

421

Semi-natural grassland

1113

Low Density Fabric (IM.D <30%)

422

Alpine and sub-alpine natural grassland

112

Industrial, commercial and military units

511

Heathland and Moorland

121

Road networks and associated land

512

Other scrub land

122

Railways and associated land

52

Sclerophyllous vegetation

123

Port areas and associated land

61

Sparsely vegetated areas

124

Airports and associated land

621

Beaches and dunes

131

Mineral extraction, dump and construction sites

622

River banks

132

Land without current use

631

Bare rocks and rock debris

14

Green urban, sports and leisure facilities

632

Burnt areas (except burnt forest)

211

Arable land

633

Glaciers and perpetual snow

212

Greenhouses

71

Inland marshes

221

Vineyards, fruit trees and berry plantations

721

Exploited peat bog

222

Olive groves

722

Unexploited peat bog

231

Annual crops associated with permanent crops

811

Coastal salt marshes

232

Complex cultivation patterns

812

Salines

233

Land principally occupied by agriculture with significant areas of natural vegetation

813

Intertidal flats

234

Agro-forestry

821

Coastal lagoons

311

Natural and semi-natural broadleaved forest

822

Estuaries

312

Highly artificial broadleaved plantations

911

Interconnected water courses

321

Natural and semi-natural coniferous forest

912

Highly modified water courses and canals

322

Highly artificial coniferous plantations

913

Separated water bodies belonging to the river system

331

Natural and semi-natural mixed forest

921

Natural water bodies

332

Highly artificial mixed plantations

922

Artificial standing water bodies

341

Transitional woodland and scrub

923

Intensively managed fish ponds

342

Lines of trees and scrub

924

Standing water bodies of extractive industrial sites

35

Damaged forest

10

Sea and ocean

Practical considerations

The map can also be consulted online in a viewer available on the download site (see link above).

9 Coastal Zones

Product

LULC general

Dates

2012, 2018

Formats

Vector

Scale

1:10,000

Minimum Mapping Unit: 0.5 ha

Minimum Mapping Width: 10 m

Thematic resolution

71 classes:

19 (a), 6 (ag), 17 (v), 9 (m), 0 (na)

Compatible legends

Urban Atlas, N2K, Riparian Zones

Extent

Coastlines of EEA member states (39 countries)

Updating

Yes

Change detection

Yes, through the change layer

Overall accuracy

Expected to be >85%

Website of reference

Website Language English, German and French

https://land.copernicus.eu/local/coastal-zones

Download site

https://land.copernicus.eu/local/coastal-zones

Availability

Format(s)

Open Access after registration

.gdb, .gpkg

Technical documentation

European Environment Agency (2021)

Other references of interest

Project

The Coastal Zones Land Cover/Land Use dataset is produced by the European Environment Agency (EEA) as part of the Copernicus Land Monitoring Service (CLMS). The dataset has been developed in collaboration with the Copernicus Marine Environment Monitoring Service (CMEMS) and representatives from the potential community of users.

It is specifically intended for monitoring coastal areas and provides an important source of information for all EU policies dealing with coastal management and maritime spatial planning.

The dataset maps, at very detailed scale, the land uses and covers in coastal areas in the 39 countries belonging to the EEA. The coastal area mapped is defined by a 10 km inland buffer zone and the Corine Land Cover (CLC) seawards buffer zone. Relevant estuaries, coastal lowlands and nature reserves that extend beyond the buffer zone have also been included.

The dataset’s classification legend has been specifically designed to fit the needs of its user community. It is based on the Mapping and Assessment of Ecosystems and their Services (MAES) ecosystem typology and makes the product compatible with other CLMS local monitoring datasets, such as Urban Atlas, Riparian Zones and N2K.

The dataset is composed of two LUC maps for the reference years 2012 and 2018, plus a change layer for the period 2012–2018. The dataset will be updated every 6 years, in accordance with the CLC production timeline.

Production method

The Coastal Zones Land Cover/Land Use dataset is produced via computer-assisted photointerpretation of very high spatial resolution (1.5–4 m) imagery from a wide variety of missions: SPOT, Pléiades, WorldView, SuperView, KOMPSat, Planet Dove, Deimos and TripleSat. A variable photointerpretation scale (1:5,000–1:10,000) was selected depending on the mapped landscape and feature characteristics. The following auxiliary datasets were also used in support of the photointerpretation process: CLC, Urban Atlas, HRL, Bing Maps and different imagery sources (DWH_MG2_CORE_01 Coverage, Sentinel-2, Landsat-8, national aerial imagery, Google Earth).

Product description

Users can download the Coastal Zones dataset in two different formats: Geodatabase and GeoPackage. Different download files are available for each year of reference (2012, 2018) as well as for the change layer (2012–2018). All downloads include the same information: layers with LUC information, a style file for their symbolization in GIS and auxiliary data.

Downloads

Coastal Zones 2018 (Geodatabase)

– Geodatabase files with Coastal Zones vector layers

– Layer style files for ArcGIS (.lyr), QGIS (.qml) and any other GIS software (.sld) (Symbology folder)

– Metadata file (.xml, .gfs) (Metadata folder)

Database

Coastal Zones 2018 (Geodatabase)

– fid: Identifier for each polygon

– ID: Unique identifier for each polygon

– DU:

– CODE_1_18: LUC category for the Level 1 classification legend

– CODE_2_18: LUC category for the Level 2 classification legend

– CODE_3_18: LUC category for the Level 3 classification legend

– CODE_4_18: LUC category for the Level 4 classification legend

– CODE_5_18: LUC category for the Level 5 classification legend

– COMMENT_18: Comments on the mapping

– NODATA_18: Objects with no data in 2018

– AREA_HA: Area of the polygon, in hectares

– Shape_Length: Perimeter of the polygon, in metres

– Shape_Area: Area of the polygon, in square metres

Coastal Zones change 2012–2018 (Geodatabase)

– fid: Identifier for each polygon

– ID: Unique identifier for each polygon

– DU:

– CODE_1_12: LUC category for the Level 1 classification legend in 2012

– CODE_2_12: LUC category for the Level 2 classification legend in 2012

– CODE_3_12: LUC category for the Level 3 classification legend in 2012

– CODE_4_12: LUC category for the Level 4 classification legend in 2012

– CODE_5_12: LUC category for the Level 5 classification legend in 2012

– CODE_1_18: LUC category for the Level 1 classification legend in 2018

– CODE_2_18: LUC category for the Level 2 classification legend in 2018

– CODE_3_18: LUC category for the Level 3 classification legend in 2018

– CODE_4_18: LUC category for the Level 4 classification legend in 2018

– CODE_5_18: LUC category for the Level 5 classification legend in 2018

– COMMENT: Comments on the mapping

– NODATA_12: Objects with no data in 2012

– NODATA_18: Objects with no data in 2018

– AREA_HA: Area of the polygon, in hectares

– Shape_Length: Perimeter of the polygon, in metres

– Shape_Area: Area of the polygon, in square metres

Legend and codification

The Coastal Zones dataset was produced following a hierarchical classification legend made up of five different levels, the most detailed of which is provided here. Information about the full classification scheme, including the five different levels, can be found in the technical documentation accompanying the dataset.

Code

Label

Code

Label

11110

Continuous urban fabric (IMD ≥ 80%)

36000

Damaged forest

11120

Dense urban fabric (IMD ≥ 30–80%)

41000

Managed grassland

11130

Low-density fabric (IMD < 30%)

42100

Semi-natural grassland

11210

Industrial, commercial, public and military units (other)

42200

Alpine and sub-alpine natural grassland

11220

Nuclear energy plants and associated land

51000

Heathland and moorland

12100

Road networks and associated land

52000

Alpine scrub land

12200

Railways and associated land

53000

Sclerophyllous scrubs

12310

Cargo port

61100

Sparse vegetation on sands

12320

Passenger port

61200

Sparse vegetation on rocks

12330

Fishing port

62111

Sandy beaches

12340

Naval port

62112

Shingle beaches

12350

Marinas

62120

Dunes

12360

Local multi-functional harbours

62200

River banks

12370

Shipyards

63110

Bare rocks and outcrops

12400

Airports and associated land

63120

Coastal cliffs

13110

Mineral extraction sites

63200

Burnt areas (except burnt forest)

13120

Dump sites

63300

Glaciers and perpetual snow

13130

Construction sites

71100

Inland marshes

13200

Land without current use

71210

Exploited peat bogs

14000

Green urban, sports and leisure facilities

71220

Unexploited peat bogs

21100

Arable irrigated and non-irrigated land

72100

Salt marshes

21200

Greenhouses

72200

Salines

22100

Vineyards, fruit trees and berry plantations

72300

Intertidal flats

22200

Olive groves

81100

Natural & semi-natural water courses

23100

Annual crops associated with permanent crops

81200

Highly modified water courses and canals

23200

Complex cultivation patterns

81300

Seasonally connected water courses (oxbows)

23300

Land principally occupied by agriculture with significant areas of natural vegetation

82100

Natural lakes

23400

Agro-forestry

82200

Reservoirs

31100

Natural & semi-natural broadleaved forest

82300

Aquaculture ponds

31200

Highly artificial broadleaved plantations

82400

Standing water bodies of extractive industrial sites

32100

Natural & semi-natural coniferous forest

83100

Lagoons

32200

Highly artificial coniferous plantations

83200

Estuaries

33100

Natural & semi-natural mixed forest

83300

Marine inlets and fjords

33200

Highly artificial mixed plantations

84100

Open sea

34000

Transitional woodland and scrub

84200

Coastal waters

35000

Lines of trees and scrub

  

Practical considerations

Coastal Zones files are very heavy (above 3gb), which means that the dataset may be difficult to use for those without powerful computers. The dataset can also be consulted online using the viewers available when downloading the different layers.

10 S2GLC 2017—Sentinel-2 Global Land Cover 2017

Product

LULC general

Dates

2017

Formats

Raster

Pixel size

10 m

Thematic resolution

13 classes:

1 (a), 2 (ag), 5 (v), 0 (m), 0 (na)

Compatible legends

CLC

Extent

Europe, except Russia

Updating

Not expected

Change detection

No (only one date)

Overall accuracy

Expected to be >86%

Website of reference

Website Language English

http://s2glc.cbk.waw.pl/extension

Download sites

http://s2glc.cbk.waw.pl/extension

https://finder.creodias.eu/

Availability

Format(s)

Open Access

.tiff

Technical documentation

Gromny et al. (2019a, b), Kukawska et al. (2017), Malinowski et al. (2019), Nowakowski et al. (2017)

Other references of interest

Project

Sentinel-2 Global Land Cover (S2GLC) was a project funded by the European Space Agency (ESA) in order to create an automatic methodology to globally map LUC at high resolution from Sentinel-2 imagery. The project was led by the Space Research Centre of the Polish Academy of Sciences (CBK PAN). Its main output is the S2GLC 2017 map.

The project was developed in two phases. In the first phase, the proposed methodology was tested in five prototype sites: Germany, Italy, China, Columbia and Namibia. In the second phase, the methodology was adjusted to map LUC for the whole of Europe, except Russia, Belarus and Ukraine.

Production method

S2GLC was obtained by classifying Sentinel-2 imagery. Each Sentinel-2 scene was individually classified using a set of multi-temporal images through a random forest classifier. Training data was automatically extracted from existing datasets, such as CORINE Land Cover. A set of probability rasters were obtained from the random forest classifier, and the class finally selected for each pixel was the one with the highest probability over the whole time series. A post-classification step was applied for those pixels with low probabilities.

Product description

S2GLC 2017 can be downloaded as a single file or in tiles. In the first case, users can choose to download the raster LUC file, either symbolized (RGB GeoTiff file) or not (GeoTiff file). Users who opt to download a tile from the map will automatically download both types of rasters.

Downloads

European land cover map (single-band file, RGB file)

– Raster file with LUC map

– TXT file with the product legend

– PDF with technical information about the product (tiles decomposition)

Single tile

– Raster file with LUC map

– Coloured raster file with LUC map

– Preview image of the product

– TXT file with the product legend

Legend and codification

Code

Label

Code

Label

62

Artificial surfaces and constructions

104

Sclerophyllous vegetation

73

Cultivated areas

105

Marshes

75

Vineyards

106

Peatbogs

82

Broadleaf tree cover

121

Natural material surfaces

83

Coniferous tree cover

123

Permanent snow-covered surfaces

102

Herbaceous vegetation

162

Water bodies

103

Moors and Heathland

  

Practical considerations

The map can be consulted online using the CREODIAS Browser application (https://browser.creodias.eu/). Single file download options involve very heavy files (8–16.2 gb), for which a powerful computer will be required.