CAM ml2pl: cdo 1.7.2 versus 1.9.5
Added by Aiko Voigt over 5 years ago
Dear all,
I am seeing some unexpected and confusing differences when using either cdo version 1.7.2 or 1.9.5 for a vertical interpolation of CAM data.
In version 1.7.2, I need to do the following chain of commands:
cdo -ml2pl,100000,92500,85000,70000,60000,50000,40000,30000,25000,20000,15000,10000,7000,5000,3000,2000,1000 -chvar,PS,aps u.nc u_plevint.nc ncatted -a bounds,lev,c,c,ilev u.nc ncap2 -O -s hyai=hyai*100000;hyam=hyam*100000; u.nc u.nc ncrcat -O -v U,PS,PHIS,hyai,hybi,hyam,hybm SOMEFILE u.nc
But in version 1.9.5, I need to ommit the ncap2 line, as otherwise I am getting very wrong results (vertically constant winds). Also, I don't need the chvar operator:
cdo -ml2pl,100000,92500,85000,70000,60000,50000,40000,30000,25000,20000,15000,10000,7000,5000,3000,2000,1000 u.nc u_plevint.nc ncatted -a bounds,lev,c,c,ilev u.nc ncrcat -O -v U,PS,PHIS,hyai,hybi,hyam,hybm SOMEFILE u.nc
Was there a change in how cdo handles CAM data? I couldn't find anything in the docu or in the forum.
Thanks, Aiko
Replies (2)
RE: CAM ml2pl: cdo 1.7.2 versus 1.9.5 - Added by Ralf Mueller over 5 years ago
there was one bugfix in 1.8.0 (#7225) related to missing interpolation of half-levels - not sure if that's related to this
cheers
ralf
RE: CAM ml2pl: cdo 1.7.2 versus 1.9.5 - Added by Aiko Voigt over 5 years ago
Hi Ralf,
That might be. cdo version 1.8.2 behaves like cdo version 1.9.5 in my tests.
Also, I realized that the commands given above are wrongly ordered, they should have been of course like this for cdo 1.7.2 (and analogously for 1.9.5):
ncrcat -O -v U,PS,PHIS,hyai,hybi,hyam,hybm SOMEFILE u.nc ncap2 -O -s hyai=hyai*100000;hyam=hyam*100000; u.nc u.nc ncatted -a bounds,lev,c,c,ilev u.nc cdo -ml2pl,100000,92500,85000,70000,60000,50000,40000,30000,25000,20000,15000,10000,7000,5000,3000,2000,1000 -chvar,PS,aps u.nc u_plevint.nc
Is the change in behavior for different cdo versions known among users?
Best, Aiko