[Wrf-users] Segmentation Fault with Nested NMM core Runs

[Thomas Raab]

Hello Mac,

I'm not familiar with Domain Wizard, but in the geogrid section of namelist.wps
you specify e_we=103 and e_sn=115 for domain 2 whereas in namelist.input you
have e_we=64 and e_sn=70. Use the same values as for WPS.

Greetings

Thomas

Zitat von mmacleod <mmacleod at scotiaweather.com>:

> Good morning.
>
> I am wrestling with a problem with the NMM core since the past two week.
> I always get a Segmentation Fault  on running  with  a nested grid.  We
> run the  same  two grids utilizing the ARW  core and it executes
> correctly every time.
>
> We are looking to operate a 4 km grid over the Maritime  Provinces of
> Canada to be driven by the NAM 12 km grib files available from NCEP and
> to nest a 1 .333 km grid within the 4 km grid covering the Annapolis
> Valley area of Nova Scotia and running for the same times as the 4 km grid.
>
> We have obtained the latest version of the Domain Wizard and use it to
> set-up the domains. We continue to get a segmentation fault error each
> time we try to run the NMM core with the 1 km grid.  It always comes on
> the initial cycle of domain 2 which is the 1 km grid.  We think this is
> because the system is placing grid 2 outside the window of grid one but
> we don't know why.  It could be the compile of the model but we need to
> know what to change to correct this problem if it is the compile.
>
> What we find is the domain two appears outside grid 1 when we reopen the
> Domain Wizard, even though the parameters  for each grid remain
> unchanged in the specs portion of the domain window in the Domain
> Wizard.  I have tried editing the namelist.wps and namelist.input files
> to get a match but the problem persists.
>
> I an attaching the namelist.wps, namelist.input files used to run the
> WRF_NMM core.  I am also attaching three png files that show the initial
> grid setup (HGTEST_Screen1.png), the redisplay of the grid locations
> when we open the Domain Wizard again after running the model
> (HGTEST_Screen2.png) and the changed grid with the nesting specs
> (HGTEST_Scree3.png) where one can see the nesting specs are unchanged
> despite the change in location of the windows on the Domain Wizard
> display.  I am also including a display of the  output form the
> WRF_NMM  run  showing the  Segmentation fault error.
>
> We are running on a dual Quad processor system (Xeon 5472 processors at
> 3 GHz), with 8 GBytes of RAM utilizing the Ubuntu O/S and the GNU
> gfortran compiler to build the model core.
>
> If anyone can provide guidance on the solving of our problem, we will be
> appreciative of the information.
>
> Thank you in advance for your attention to this important matter.
>
> Sincerely
>
> Mac MacLeod
>
>
>
>
> --
> M.A. (Mac) MacLeod
> President and General Manager
> Scotia Weather Services Inc
> 192 Wyse Road, Suite 8,
> Dartmouth, N.S. B3A 1M9
>
> Tele: 902-468-3866
> Fax:  902-461-1768
> E-mail: mmacleod at scotiaweather.com
> Visit us: www.scotiaweather.com
>
>