Potential Remapping bug (generic to regular)
Added by Ryan Love about 7 years ago
Hello,
I'm trying to remap a generic grid (1 face of a cubed sphere) to a regular grid and am consistently running into issues (v1.6-v1.8) where the remapping partially fails. Hoping someone can show me where I've gone wrong either in my input or similar. The mode of failure depends upon the type of remapping. Remapbil/bic will remap about 1/2 of the data properly before writing junk, remapnn/dist doesn't seem to behave well at all and writes a mix of useful data and garbage. The commands I'm using are listed below and the input and output data is attached. 4 of the other faces remap fine, one other fails in a similar manner so I suspect something is similarly wrong.
REMAP_EXTRAPOLATE=off cdo remapnn,r360x360 side_5.l1.nc side_5.remapnn.360x360.l1.nc
REMAP_EXTRAPOLATE=off cdo remapdis,r360x360 side_5.l1.nc side_5.remapdis.360x360.l1.nc
REMAP_EXTRAPOLATE=off cdo remapbil,r360x360 side_5.l1.nc side_5.remapbil.360x360.l1.nc
REMAP_EXTRAPOLATE=off cdo remapbic,r360x360 side_5.l1.nc side_5.remapbic.360x360.l1.nc
Cheers and thanks,
Ryan
Replies (3)
RE: Potential Remapping bug (generic to regular) - Added by Ryan Love about 7 years ago
sanitized cdo -V output below in case reference info is needed:
Climate Data Operators version 1.8.0rc5 (http://mpimet.mpg.de/cdo)
Compiled: by rlove on X (x86_64-unknown-linux-gnu) Jan 4 2017 11:10:26
Compiler: gcc -g -O2 -fopenmp
version: gcc (GCC) 6.2.1 20160916 (Red Hat 6.2.1-2)
Features: DATA PTHREADS OpenMP45 HDF5 NC4/HDF5 SSE2
Libraries: HDF5/1.8.16
Filetypes: srv ext ieg grb1 nc1 nc2 nc4 nc4c
CDI library version : 1.8.0rc5 of Jan 4 2017 11:09:37
CGRIBEX library version : 1.7.6 of Nov 5 2016 11:59:02
NetCDF library version : 4.4.0 of Feb 4 2016 12:10:14 $
HDF5 library version : library undefined
SERVICE library version : 1.4.0 of Jan 4 2017 11:09:26
EXTRA library version : 1.4.0 of Jan 4 2017 11:09:21
IEG library version : 1.4.0 of Jan 4 2017 11:09:24
FILE library version : 1.8.2 of Jan 4 2017 11:09:22
RE: Potential Remapping bug (generic to regular) - Added by Uwe Schulzweida about 7 years ago
Thanks for this report! I have created a bug report out of this post, see bug #7625.
RE: Potential Remapping bug (generic to regular) - Added by Ryan Love about 7 years ago
Sounds good, hopefully nothing too tricky to sort out. If more information would be helpful do let me know. I will add (for additional information, and for those who may find similar issues down the road) that when I merged all the faces into one single netCDF file (by stacking the faces along the y-dimension) that the remapping was markedly successful despite the rather odd looking input.