next
previous
items

Skip to content. | Skip to navigation

Sound and independent information
on the environment

– or –
Please login or register to participate.
Image

Errors in Corine Land Cover 2000 Version 15?

.
Description
Hi!
I am working with some data of the metropolitan area of Manchester and just noticed that there were some very strange land conversions in the area from 2000 to 2006 recorded, when looking on the latest CORINE data (Version 15). A major urban area west of Manchester disappeared in CLC 2006, a big lake appeared north of it and agriculture changed quite a lot. However, when looking into the separate change file which can be downloaded from your database, these changes are not visible.

The figure shows the problem. I was wondering on which source the change file is based instead?

Christian
Comments (2)
EEA Nov 17, 2011 02:19 PM
Dear Christian,

Thank you for your enquiry.

According to CLC methodology a direct comparison of 2 status layers (2000 and 2006) should be avoided (due to 5 ha minimum mapping units (MMU) in change and 25 ha in status layer).

If an analysis of change is needed CLC change layer should be applied. Theoretically, CLC2000 + CLC00-06 changes = CLC2006 (spatial inconsistency might occur due to mentioned 5/25 ha MMUs).

1. CLC2000 is taken as a reference layer to compare with IMAGE 2006 and only changes bigger than 5 ha are recorded in a separate CLC2000-2006 change layer.

2. CLC change layer is merged with CLC2000 and only minimum 25 ha polygons are maintained, thus creates CLC2006 (separate polygons less than 25 ha are dissolved).

However, the UK CLC 2000, 2006 and change 2000-2006 layers were all produced separately thus they do not correspond fully to each other:

1) CLC2006 UK has only a limited connection to CLC2000. In most of the other countries CLC2006 was produced by adding together CLC2000 and CLC-Changes(2000,2006). CLC2006-UK has been produced by a methodology, which is not the same as CLC2000 methodology. Consequently CLC2000 and CLC2006 might be strongly different and cannot be compared in the UK to derive changes.

2) CLC2006 and CLC-Changes 2000-2006 were produced independently in the UK, which raised serious thematic and topologic problems, i.e. frequently CLC-Changes 2000-2006 were in contradiction with CLC2006.

Accordingly to the CLC verification, we can conclude that CLC2006 and CLC-Changes 2000-2006 are more realistic than CLC2000. But be aware that it might not be 100% valid. We suggest to compare the datasets with IMAGE2000 and IMAGE2006.

Another option is to ask for UK national land cover data from the national authorities in the UK. This might serve your needs better. The contact details for the EEA National Reference Centre for Land Cover in the United Kingdom can be found here:
http://eea.eionet.europa.eu[…]&af=0&ud=1&od=1

We hope this information is useful to you.
landblend Nov 17, 2011 02:33 PM
Thanks for your detailed answer!

I was aware of the minimum mapping units, but these differences striked me anyway. And I did not know that is was so problematic to compare 2 status layers. Also because usally they work relatively well.

I think I will try to get data from the UK land cover (http://www.ceh.ac.uk/LandCoverMapping.html)
Thanks again.
 
Loading