European Commission logo
INSPIRE Community Forum

Copernicus RDA EU-DEM EGCS naming

Happy New Year everyone!

I have been using the Copernicus EU-DEM product a few months now but only realised recently (when I wanted to automate some processing) that the European Grid Coordinate System (EGCS) naming used for the EU-DEM tiles is not INSPIRE compliant. Can somebody confirm this?

According to INSPIRE rules, the cell origin should be lower left when using the EGCS. However, the EU-DEM is using the centre pixel for naming the tile. Am I confused? Or should the EEA fix the naming? Any references that help better understand this would also be much appreciated.

Since I'm developing a downstream product, this is a bit of an issue. The reason is that you want your customer to download the correct tile based on the EGCS naming. Otherwise, they could potentially download the wrong tile. However, they would quickly realise their (or my) mistake because when they would try and overlay the data with other geographic data, there would be a shift.

Example EU-DEM tile filename: EUD_CP-DEMS_4500025000-AA.tif (level 5 for 25m resolution)

The output of gdalinfo:

Upper Left  ( 4000000.000, 3000000.000) (  5d31' 4.07"E, 50d 1'26.83"N)
Lower Left  ( 4000000.000, 2000000.000) (  6d11'52.97"E, 41d 1'36.49"N)
Upper Right ( 5000000.000, 3000000.000) ( 19d26'39.40"E, 49d43' 8.38"N)
Lower Right ( 5000000.000, 2000000.000) ( 18d 1'22.97"E, 40d46'27.19"N)
Center      ( 4500000.000, 2500000.000) ( 12d17'27.23"E, 45d35'15.84"N)

Notice the name is based on the Center pixel and not the Lower Left pixel.

Thanks in advance and have a great day,




  • Jordi ESCRIU

    Dear Conrad,

    Happy New Year 2015!

    I assume you are working with EU-DEM in ETRS89-LAEA projection (EPSG code 3035) and you are referring to the "European Grid Coding System (EGCS)". Am I right?

    Does anyone know where the EGCS is formally described?

    I am just downloading a EU-DEM tile to check the corresponding data and metadata, but I would say that the EU-DEM grid is not totally conformant to the INSPIRE rules.

    Apparently, the reason may be that the EU-DEM was produced before the stablishment of the INSPIRE rules - in 2010 concerning the first definition of the "Grid_ETRS89-LAEA", namely "Equal Area Grid" - (Commission Regulation (EU) No 1089/2010 of 23 November 2010 as regards interoperability of spatial data sets and services), also with amendments later on.

    Note that all grids based on ETRS89-LAEA (the projected CRS) do not comply with "Grid_ETRS89-LAEA" (e.g. the 25 m grid resolution in your example, is not foreseen in the resolution levels of the "Grid_ETRS89-LAEA").

    I will come back to you later on when having more information.

    Many thanks for your post!


  • Conrad BIELSKI

    By Conrad BIELSKI

    Hi Jordi,
    thanks for the clarification. I must admit that I haven't looked at the EGCS in quite sometime and there is a lot of 'old' information still floating about. I certainly agree that just using EPSG 3035 is not enough to be INSPIRE compliant.

    From the documentation you link to, it seems that INSPIRE compliant datasets are the following:

    The grid is hierarchical, with resolutions of 1m, 10m, 100m, 1 000m, 10 000m and 100 000m.

    This certainly helps me a lot, thanks!

    In the case of the EU-DEM, I think that if they were to make it INSPIRE compliant, coarsening would not be a problem and could be accomplished with an automated procedure. Producing a 10m EU-DEM would be great as long as it would be created from the correct scale of observation as input data.

    Thanks again,


  • Peter STROBL

    Happy New Year also from my side!

    A short comment on the 'EGCS' of which I happen to be one of the originators. It goes back to 2005 when we saw a need for an (at least) JRC internal standard for raster data encoding. It was mainly based on the "Reference grids for Europe" report (Annonni et al 2005).

    The basic ideas were:

    ETRS89/ETRS-LAEA (EPSG3035) as projection

    19 quadtree organised grid levels from 1m-1000km (i.e. 1,2.5,5,10,25,50,100,...m)

    a machine-readable coding system which assign each valid cell at each valid level a unique and as simple as possible code

    The underlying ideas were given up when the regulation was drafted, so it's true that today it is  non-compliant to the INSPIRE D2.8I2 Equal Area Grid. But keep in mind that the latter “…is intended more for statistical reporting purposes. It is not intended for communities where the grids have to be optimised for data exchange, supercomputer processing and high volume archiving of new data each day.”

    Instead for reference data a Zoned Geographic Grid is recommended, but that's a different story.


  • Peter STROBL

    After having checked a bit more in detail the elevation specifications this is to confirm that the currently distributed EUDEM version is NOT INSPIRE compliant in at least the following points:

    - EPSG3035 is NOT an allowed projection (and 25m is NOT an allowed grid spacing)

    - cell origin used for the coding is the lower left corner (and NOT the upper left)

    - ECGS cell codes are NOT INSPIRE compliant

    - interpolation used for resampling was hermitic spline NOT bilinear

    - vertical accuracy and resolution are not within the specified relation margins

    So, to make a long story short: The whole data set would have to be redone from scratch. But I have doubts that this is even worth it, given that with the release of the 1" SRTM data, large parts of EUDEM can be considered outdated.

  • Conrad BIELSKI

    By Conrad BIELSKI

    Hi Peter,

    thanks for the link and confirmation. I think that such forums are important to push forward these issues. Therefore, in a perfect world, we would have access to sub-10m height data to produce a EU wide INSPIRE compliant DEM to share. In such a case, what would be the best projection, vertical accuracy, and naming scheme?

    Have a great day,


  • Peter STROBL

    you name it Conrad - in a perfect world!

    Till then we'll have to cope with what they call reality - even if it's virtual ;)

    What regards your question, I've just started familiarising myself with these guidelines, but my understanding is the following:

    Data would come as Zoned Geographic Grid in ETRS89_GRS80 which means:

    - No projection and a grid with variable resolution in longitudinal direction surprise

    - vertical accuracy is tied to horizontal resolution

    - coding would follow INSPIRE rules (e.g. Grid_ETRS89-GRS80z2_1S for the 1" grid)

    While this of course avoids any issues with projections like angles, areas, directions, etc, it leaves it pretty open how one should store and work with these data (without projecting them first).

    Therefore my question:

    - Does anyone out there have a pan-European data set that fulfills these criteria? Which SW do you use to create and handle it?


  • Jordi ESCRIU

    Many thanks, Peter and Conrad!

    I think it is clear that the current version EU-DEM is not INSPIRE compliant and this dataset may be considered outdated for certain applications and purposes.

    In my opinion the EU-DEM is a pan-European product owned / ordered by the Directorate-General Enterprise and Industry (DG-ENTR) / European Commission (EC) which was produced by the private sector. Therefore, the product itself is harmonized at European level (no contributions from Member States) and ready to use for appropriate purposes.

    I would say that the objective of INSPIRE is to harmonize existing data in Member States (more accurate and updated), establishing general and thematic rules to share this data in an interoperable way and use it for the required EU and/or crossborder purposes.

    Some remarks:

    • EPSG 3035 (ETRS89-LAEA) is one of the two-dimensional / plane CRSs allowed by INSPIRE for making data available. Although this CRS is recommended for spatial analysis and reporting (where true area representation is required), it may be used for making data available for INSPIRE.
    • The EL data specification allows data interpolated by using other methods different than bilinear (from those foreseen within the CV_InterpolationMethod code list in ISO 19123). The bilinear method is just recommended and identified as a default value.
    • The relation margins provided in the EL data specification between vertical accuracy and resolution (grid spacing) are just recommended. Data according other relation margins is accepted.

    I feel the present discussion topic mix EU-DEM usability aspects (related to the EL theme or sub-group) together with grid use / usability aspects (GG theme or sub-group).

    Please use the following newly open threads in the corresponding subgroups for further discussion, if required:

    NOTE: In order to participate in such discussions, you must join to the corresponding sub-groups.

    Feel free to open a new discussion topic, if not concerned with the previous ones.

    Have a nice discussion!

This discussion is closed.

This discussion is closed and is not accepting new comments.

Elevation, Ortho & Grids

Elevation, Ortho & Grids

INSPIRE Thematic Cluster Elevation, Orthoimagery, Reference systems, Geographical grids - Join this group to share your knowkledge, learn and collaborate in solving issues related to the Elevation, Orthoimagery, Reference systems and Geographical grids themes