[Wrf-users] Upper boundary cfl error (UNCLASSIFIED)

[Thomas Raab]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

I also ran a couple of simulations with an innermost grid distance of
300m using 1" topograpphy data. First I found the same issues with rapid
  instability growth at a single grid point at steep slopes, regardless
of model time step. I then shifted the innermost grid a little bit
around so that after some trial and error I found a position where the
model did not blow up. Seems to be there is a crucial steepness  above
which the model gets unstable. Height of first vertical level above
ground and number of vertical levels was also important.

I hope this is acknowledged in future releases, at the moment WRF is not
really suitable for very high resolution simulations.

Best regards,

Thomas


Dumais, Bob (Civ, ARL/CISD) wrote:
> Classification: UNCLASSIFIED
> Caveats: NONE
> 
> Thank you , Mark. That is a good suggestion, especially as if you say
> the hi res structure stays intact even after 5 passes. I would be
> worried if the 5 pass smoothing made 500 m terrain field look more like
> a 2 km terrain field after 1 pass. As Sen Chiao mentioned earlier,
> without capturing the higher resolution terrain & land use structure,
> the impact of a higher resolution grid becomes questionable in most
> cases when considering the computational cost involved. Have a great
> night-
> 
>                                                      Bob
> 
> -----Original Message-----
> From: wrf-users-bounces at ucar.edu [mailto:wrf-users-bounces at ucar.edu] On
> Behalf Of Mark Stoelinga
> Sent: Thursday, March 04, 2010 3:50 PM
> To: wrf-users at ucar.edu
> Subject: Re: [Wrf-users] Upper boundary cfl error (UNCLASSIFIED)
> 
> I can also confirm what Bob and David suggested...I had the same problem
> for 4/3-km simulations, and solved it the same way ( don't recall if we
> did 4 or 5 smoothing passes).
> 
> What I do recall is that the resultant smoothing of the terrain is not
> as drastic as one might fear from 5 smoothing passes.  The differences
> are noticeable, but most of the hi-res structure in the terrain is still
> there.  And the WRF crash problem goes away.
> 
> Mark
> 
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v2.0.14 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org

iEYEARECAAYFAkuQ5F8ACgkQPfTWv/pc8VuNLACeO8SAXiSkQ1feUK2GKRO1ZQh/
rUQAnjwURc9Zm7d5KebPuKlRNkgEYInd
=yKNY
-----END PGP SIGNATURE-----