Forums » Open discussion »
remapnn file or method corrupt?
Added by Swantje Preuschmann over 11 years ago
Dear Ralf, Uwe
greating from Chilehouse Hamburg.
I am logically a bit astonished by one remapnn result. So, before I am going into details I would like to ask you quickly if you are able to tell me, why remapnn is doing such a weird structure.
What I have done in short:
- original file: a tile in roughly 0.001 deg horizontal resolution
- cdo remapcon,EUgrid0.5deg -o rmcon5
- cdo remapbil,EUgrid0.5deg -o rmbil5
- cdo remapnn,EUgrid0.5deg -o rmnn5
rmcon5:
Sorry.... send it by accident, while trying to see a picture ....
Now the point:
as you can see in rmnn5 the original tile exists as in rmbil and rmcon, but has a wide extinction to the left.
There it should also be missing values in my understanding. Somehow, the boundary boxes are not
used correctly, even though, all fields in the original file have exact boundary boxes (as attached as originalTile).
Hi Swantje,
this is the result of the remapping extrapolation feature. This feature is enable by default for remapnn and remapdis. You can switch it off with the environment variable REMAP_EXTRAPOLATE=off.
Cheers,
Uwe
JUCHEEEE!!!
Thank you!!!
:-)))))