[RegCNET] Error in wind field preprocessed by ICBC, when NORMER is used
Graziano Giuliani
ggiulian at ictp.it
Fri Jun 29 11:29:02 CEST 2012
On 06/28/2012 03:22 PM, Juan Bazo wrote:
> Hi Felipe,
>
> I found the same problem with ReCM 4.1.1 and 4.2, now I use ROTMER
> projection, because i don't find any solutions for this bug. Also I have
> pics for two situations.
>
> Regards from Peru.
>
[...]
Just to be sure that this is to be investigated again, and that it is not a an
already examined issue, are those plots obtained with grads?
We are aware of a bug in grads when opening two files containing packed netcdf
datasets (integer coded, with add_offset and scale_factor attributes).
The following procedure leads to WRONG results:
open file 1
read u
close 1
open file 2
read v
close 2
plot u,v
In this case grads is using the scale_factor and add_offset of the last read
variable (v in this case) to calculate the value also of the first variable.
The resulting vector components are wrong and the plot is not representing the
field. This can solved placing both wind components in a single file and
plotting it:
open file 1
plot u,v
this pseudo code works, the previous does not. So to be clear I am again here
listing some remarks and asking some questions.
Remarks:
1) A post with just the message "it doesn't work for me" without posting at
least the relevant bits from the regcm.in is useless for us to investigate.
2) On reporting, always specify which analysis tool have been used to produce
graphical results (ferret,NCL,R,python,matlab,idl,grads..)
Questions:
1) Can one of the people reporting this issue please post the relevant bits from
the namelist input file? (dimparam,geoparam,terrainparam,globdatparam)
2) Which graphical tool has been used to spot the issue and which procedure have
been followed to compare ICBC and global model data?
Thanks,
Graziano.
--
Graziano Giuliani - Earth System Physics Section
The Abdus Salam International Centre for Theoretical Physics
Strada Costiera, 11 - I - 34151 Trieste Italy
More information about the RegCNET
mailing list