[Wrf-users] WRF 3.2 Model not running after 4hrs of Integration.

Kevin Matthew Nuss wrf at nusculus.com
Wed Mar 9 12:23:11 MST 2011


Hi Geeta,

I noticed that in your rsl.error.0003 file you had multiple CFL errors.
These were kind of hidden because you did not delete the rsl file between
running real.exe and wrf.exe.

When CFL errors occur, I often get segmentation faults, usually because the
radiation scheme. Segmentation faults can cause core dumps. If not caused by
the radiation scheme, maybe it occured in the Kains Fritsch cumulus scheme.
I say that because you did get the message: "WOULD GO OFF TOP: KF_ETA_PARA
I,J,DPTHMX,DPMIN 40 18 NaNQ 5000.000000" which includes a NaN.

To fix the CFL error, you may have to keep reducing the timestep, even
though yours is not long for your grid resolution. That may fix things.
If the timestep becomes impractically short, try adding some vertical
dampening:
  w_damping = 1
or play around a little with epssm:
  epssm = 0.3, 0.3, 0.3
Increasing the vertical size of grid cells might help too with CFL errors,
but you probably have those where you want them.

It will help you find the errors I found if you delete or remove the rsl
files before running wrf.exe.

best wishes,
Kevin

On Thu, Mar 3, 2011 at 9:32 PM, Geeta Geeta <geeta124 at hotmail.com> wrote:

>  Dear Sir,
> Thanks for the reply. I am running WRf 3.2 for the 3 domains 27, 9 and 3 km
> for the 48hrs Forecast on the IBM P570 Machine. The Operating system is
> AIX5.3 and the code is compiled using xlf with POE in place.
> I had given the command *bash$ time poe ../wrf.exe -procs 8. *
>
> 1. The model gave segmentation fault after 4hrs of Integration. I also
> plotted the Model forecast for 2nd and 3rd hour at 27km resolution (Domain
> 1) So it showed me some data being plotted
> *But when I tried to plot the output for the 2nd and 3rd domain, IT says ENTIRE
> GRID UNDEFINED. Kindly suggest.  *
>
> 2. >>>What data I have used ??
> *I am using the .grib2 files as Initial and BC taken from
> ftp://ftpprd.ncep.noaa.gov.in at 1x1 degree resolution.
> *
> 3. >>>What is the error message in your output file? Did you find any CFL
> violation?  Your namelist.input looks fine. Please send us error messages in
> your run.
> *I am attatching the rsl.error* and the rsl.out* files, namelist.input and
> namelist.wps files. It seems to me that there is no CFL violation. *
>
> Thanks.
> geeta
>
>
>
> ------------------------------
> Date: Thu, 3 Mar 2011 17:20:41 -0700
> From: wrfhelp at ucar.edu
> To: geeta124 at hotmail.com
> Subject: Re: WRF 3.2 Model not running after 4hrs of Integration.
>
> Geeta,
>
> What is the error message in your output file? Did you find any CFL
> violation?
>
> Your namelist.input looks fine. Please send us error messages in your run.
>
>
> On 3/3/2011 3:57 AM, Geeta Geeta wrote:
>
> Dear All,
> I am running WRF3.2 for 3 domains, at 27, 9 and 3kms.  After integrating
> for 4hours, the model  gives segmentation fault. and these directories are
> created and does not integrate beyond.
>
> bash-3.2$ ncdump -v Times wrfout_d01_2010-08-04_00:00:00
> Times =
>   "2010-08-04_00:00:00",
>   "2010-08-04_01:00:00",
>   "2010-08-04_02:00:00",
>   "2010-08-04_03:00:00",
>   "2010-08-04_04:00:00" ;
> }
>
> bash-3.2$ ncdump -v Times wrfout_d02_2010-08-04_00:00:00
> Times =
>   "2010-08-04_00:00:00",
>   "2010-08-04_01:00:00",
>   "2010-08-04_02:00:00",
>   "2010-08-04_03:00:00",
>   "2010-08-04_04:00:00" ;
>
> bash-3.2$ ncdump -v Times wrfout_d03_2010-08-04_00:00:00
> Times =
>   "2010-08-04_00:00:00",
>   "2010-08-04_01:00:00",
>   "2010-08-04_02:00:00",
>   "2010-08-04_03:00:00",
>   "2010-08-04_04:00:00" ;
>
> After integrating for 4hours, the model  gives segmentation fault. and
> these directories are created and does not integrate beyond.
> drwxr-xr-x    2            256 Mar 03 04:20 coredir.5
> drwxr-xr-x    2            256 Mar 03 04:20 coredir.3
> drwxr-xr-x    2            256 Mar 03 04:20 coredir.2
> drwxr-xr-x    2            256 Mar 03 04:20 coredir.1
>
> Kindly help. The namelist.input file is also attached for reference.
>
> Geeta
>
>
>
> _______________________________________________
> Wrf-users mailing list
> Wrf-users at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/wrf-users
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ucar.edu/pipermail/wrf-users/attachments/20110309/3969bd7d/attachment.html 


More information about the Wrf-users mailing list