Period: 2012, 2015

Sensor: 2012 – IRS-p6, ResourceSat-2, SPOT-4, SPOT-5 (for 2015 +/- 1year)

2015 : Sentinel 2, Landsat 8 (for 2015 +/- 1year)

Resolutions: 20 and 100m

Extent: Pan-European

Data can be accessed here.

Algorithm description: The product comprises of three datasets: 1. Tree cover density providing information about tree cover density (0 to 100%), 2. Dominant leaf type (broadleaf/coniferous), 3. Forest type product which has been developed so as to be consistent with FAO forest types. Forest change product comprises of simple tree cover density change for 2012-2015.

Classification algorithm is semi-automatic wherein high resolution satellite images are subjected to scene based supervised and unsupervised classification. Time series analysis is performed to get dominant leaf type and tree cover density. This is followed by interactive manual corrections in the tree cover map.

Accuracy: Accuracy assessment for entire Europe has not yet been done for this product. Sarmento et al (2015) report results for Portugal and they found that the overall accuracy for forest type map was 79.7% with an absolute precision of 2.2% at the 95% confidence level. In general, there was overestimation of non-forest class resulting in underestimations in forest class. This also has impact on the accuracy for different forest types. The user’s  (producer’s) accuracy for broadleaf, coniferous and mixed forest were 73% (56.3%), 64% (41.4%) and 29.5% (24.8%) respectively.

Congedo et al (2016) found that the HRL data are of ‘acceptable’ accuracy (85% overall accuracy) in case of most of the verified areas in Italy. However they promote country/region specific enhancement of the maps to increase accuracy.

Limitations:

  • Limited country-specific information is available for verification.

Reference:

Sarmento, P., Marcelino, F., Monteiro, G., Schmedtmann, J. & Caetano, M. Accuracy assessment of Copernicus program 2012 High-Resolution Layers for Continetal Portugal. 1–36 (2015).

Congedo, L. et al. Copernicus high-resolution layers for land cover classification in Italy. J. Maps 12, 1195–1205 (2016).