issue with remapcon and missing values
Added by Patrick Jöckel over 7 years ago
Dear Colleagues,
I have a problem with the remapcon operator of cdo (1.8.0 and 1.7.1):
Whereas the result of
cdo -b 64 -O -f nc setmisstoc,0.0 -remapcon,T42L90MA_gp.grb sample1.nc sample1_out.nc
gives the expected results, the same command with a very similar file
cdo -b 64 -O -f nc setmisstoc,0.0 -remapcon,T42L90MA_gp.grb sample2.nc sample2_out.nc
yields a strange pattern in the output data (see compare_output.png).
Does anybody have an idea and can help?
compare_output.png (35.9 KB) compare_output.png | |||
sample1.nc (403 KB) sample1.nc | |||
sample2.nc (403 KB) sample2.nc | |||
T42L90MA_gp.grb (24.1 KB) T42L90MA_gp.grb |
Replies (3)
RE: issue with remapcon and missing values - Added by Ralf Mueller over 7 years ago
Hi Patrick!
I cannot reproduce this with 1.8.1. I created both plot with 'nearest' interpolation of matplotlib - so any local variation should be visible. but both look ok IMO
Differences are visible, but now strange patterns like in your plots. Could you test it with 1.8.1, too?
hth
ralf
sample2_out.nc.png (33.9 KB) sample2_out.nc.png | |||
sample1_out.nc.png (32.8 KB) sample1_out.nc.png |
RE: issue with remapcon and missing values - Added by Ralf Mueller over 7 years ago
1.8.0 was taken from the download area because we encountered an interpolation related bug. I'm not sure, if this one is hitting you, but updating to 1.8.1 is highly recommended.
RE: issue with remapcon and missing values - Added by Patrick Jöckel over 7 years ago
Dear Ralf,
indeed the issue is gone with 1.8.1 of cdo.
Thank you very much!!!
Yours,
Patrick