Subject: Re: modifying initial PHIS data set
From: Jim Rosinski (rosinski@cgd.ucar.edu)
Date: Wed Nov 17 1999 - 12:58:37 MST
On Tue, 16 Nov 1999, Christopher Poulsen wrote:
> I would like to modify the initial CCM3 PHIS data set. It appears from
> examining the SEP1 PHIS that it is spectrally transformed. My question is
> whether this spectral transformation is necessary in the creation of a
> new PHIS data set or whether raw PHIS values will work as an initial
> condition.
Even if you input raw (non-truncated) PHIS values, the model will spectrally
truncate the field on startup. The reason it does this is that the surface
geopotential appears in the dynamical equations, and therefore the truncated
PHIS is what the model actually "sees".
Contrary to an earlier message, you should NOT add variable-specific netcdf
attribute "from_hires" to your initial dataset unless you are sure of the
need for it. The presence of this attribute has to do with a spectral filter
that needs to be applied when the surface geopotential has been generated
from a high resolution topographic dataset. It is unrelated to spectral
truncation, which will happen no matter what when the model is run. This
attribute should only exist and be true when the offline program "definesurf"
has been run to produce the PHIS field on the initial dataset. That program
defines the attribute automatically.
Regards,
Jim Rosinski
CCM Core Group
This archive was generated by hypermail 2b27 : Thu Jun 01 2000 - 09:28:59 MDT