Project

General

Profile

remapping: bil works fine, con/con2 segmentation fault

Added by Antje Fitzner over 11 years ago

Dear all,

I want to remap climate data using conservative remapping (con and con2 give the same output). When doing the same for bilinear remapping, it works fine!!

Here is what I have been doing:
$ cdo remapcon,grid.nc in.nc out.nc # same for con2
11668 Segmentation fault (core dumped)

$ cdo gencon,grid.nc in.nc weights.nc
$ cdo remap,grid.nc,weights.nc in.nc out.nc
12718 Segmentation fault (core dumped)

Is there something I am overlooking? I am a bit confused since bil does work.

Thanks for your help!

PS: I am using CDO 1.5.3


Replies (4)

RE: remapping: bil works fine, con/con2 segmentation fault - Added by Uwe Schulzweida over 11 years ago

Hi Antje,

Could you please attach an example file? That will help us a lot to solve this problem.

Cheers, Uwe

RE: remapping: bil works fine, con/con2 segmentation fault - Added by Antje Fitzner over 11 years ago

Dear Uwe,

Sure, attached my original files in.nc and grid.nc

Thanks for your help!
Antje

RE: remapping: bil works fine, con/con2 segmentation fault - Added by Uwe Schulzweida over 11 years ago

Dear Antje,

Thanks for the files! I can reproduce this error. A "segmentation fault" is definitely a bug in CDO. I will try to solve this problem in the next release.
The reason for this bug is an unexpected structure of the arrays grid_corner_lat and grid_corner_lon. The dimension order is
"(grid_corners, rlat, rlon)", expected is "(rlat, rlon, grid_corners)". An other problem is that the units of these arrays differ from the units of the coordinates lat and lon.

Best regards,
Uwe

RE: remapping: bil works fine, con/con2 segmentation fault - Added by Antje Fitzner over 11 years ago

Dear Uwe,

Thanks for your help!
Changing the units and order solved it!

Best,
Antje

    (1-4/4)