CDO

Original from antonella.sanna@cmcc.it

using cdo with files obtained after cmor postprocessing I got these two warnings:

cdo -r -fldmean tas_Amon_CMCC-CM_decadal1960_r1i1p1_196001-196912.nc ~/pippo.nc

Warning (cdfSetDim) : Inconsistent dimension definition for height! dimid = 0; type = 0; newtype = 3

Warning (scanAttributes) : NetCDF: Variable not found - areacella

cdo fldmean: Processed 13824000 values from 1 variable over 120 timesteps. ( 1.40s )

Martin Juckes suggested to me to contact the cdo support to check if there is something wrong or the file is written in proper way.

If you are not the right person to contact for this problem please forgive me and please forward this email to the devoted person.

Many thanks for your kindness

Antonella Sanna
CMCC-Bologna
Italy

Begin forwarded message:

From: <martin.juckes@stfc.ac.uk>
Date: March 15, 2011 2:34:50 PM GMT+01:00
To: <antonella.sanna@cmcc.it>, <cmor@lists.llnl.gov>
Subject: RE: warning with cdo

Hi Antonella,

it looks to me as if, for the 2nd warning, cdo is reacting to a CF non-conformance in the file created by CMOR2: the variable indicated by the cell_measures attribute (areacella) should be in the same file according to the convention, but isn't. I'd contact cdo support to check that the code deals with this in the way you would hope - namely using the latitude and longitude specified in the dimensions to calculate the grid cell area,

regards,
Martin