[Wrf-users] RE: Wrf-users Digest, Vol 42,
Issue 14--High vertical resolution
Hjelmfelt, Mark
Mark.Hjelmfelt at sdsmt.edu
Thu Feb 28 13:38:24 MST 2008
Actually the large time steps I used for 0.5 km horizontal resolution
were a couple of seconds and the small time step was about 0.3-0.5 s
Mark Hjelmfelt
Professor and Chair, Department of Atmospheric Science
South Dakota School of Mines and Technology
501 E. St Joseph Street, Rapid City, SD 57701
605-394-1991
-----Original Message-----
From: Hjelmfelt, Mark
Sent: Thursday, February 28, 2008 12:54 PM
To: 'wrf-users at ucar.edu'
Subject: RE: Wrf-users Digest, Vol 42, Issue 14--High vertical
resolution
You do have real vertical motions in the model, these are subject to CFL
issues same as horizontal motions. For regional simulations and
moderate vertical resolution, these would be small compared to the
horizontal motions. At convective scales these may be comparable. The
CFL limits need to consider the vertical motions same as the horizontal.
In addition, unresolved turbulence can add slightly to the instability
and further reduce the usable time step slightly from what you would
calculate. In working with 3-d Cloud models we have to do this. And
ARPS, for example, has an example for doing this in their users guide.
Doing high resolution lake effect simulations I've often been forced to
a few tenths of a second time step.
Mark Hjelmfelt
Professor and Chair, Department of Atmospheric Science
South Dakota School of Mines and Technology
501 E. St Joseph Street, Rapid City, SD 57701
605-394-1991
-----Original Message-----
From: wrf-users-bounces at ucar.edu [mailto:wrf-users-bounces at ucar.edu] On
Behalf Of wrf-users-request at ucar.edu
Sent: Thursday, February 28, 2008 12:00 PM
To: wrf-users at ucar.edu
Subject: Wrf-users Digest, Vol 42, Issue 14
Send Wrf-users mailing list submissions to
wrf-users at ucar.edu
To subscribe or unsubscribe via the World Wide Web, visit
http://mailman.ucar.edu/mailman/listinfo/wrf-users
or, via email, send a message with subject or body 'help' to
wrf-users-request at ucar.edu
You can reach the person managing the list at
wrf-users-owner at ucar.edu
When replying, please edit your Subject line so it is more specific
than "Re: Contents of Wrf-users digest..."
Today's Topics:
1. Experiences with high-resolution WRF-ARW (Kemp, Eric M.)
2. Re: Experiences with high-resolution WRF-ARW
(Gustafson, William I)
3. Re: Experiences with high-resolution WRF-ARW (David Ovens)
4. FW: [Wrf-users] Experiences with high-resolution WRF-ARW
(Kemp, Eric M.)
----------------------------------------------------------------------
Message: 1
Date: Wed, 27 Feb 2008 16:01:49 -0500
From: "Kemp, Eric M." <Eric.Kemp at ngc.com>
Subject: [Wrf-users] Experiences with high-resolution WRF-ARW
To: <wrf-users at ucar.edu>
Message-ID: <52B67E1F32707847B44B4B7B1238E363020D568E at xmbv3801>
Content-Type: text/plain; charset="iso-8859-1"
Dear wrf-users:
My group is running WRF-ARW at high resolution (dx = 1 km) to simulate
turbulence.
Recently we tried running at very high vertical resolution (dz ~= 12.5 m
in the lowest
500 m of the domain), and encountered CFL errors with time steps as
low as 1/8 of a second. These runs used the 6th order diffusion
option from Knievel et al (2007), along with real terrain (area around
Albuquerque, NM) and physics (MYJ TKE, Noah land surface,
RRTM/Dudhia radiation, and WSM5 microphysics).
Has anyone here run WRF at such fine resolution, and if so, what were
your experiences with the time step? The ARW documentation recommends
6 seconds for each kilometer of dx, ignoring the effect of vertical
resolution.
Are there other settings that should be tuned?
-Eric
Eric M. Kemp
Meteorologist
Northrop Grumman Information Technology
Intelligence Group (TASC)
4801 Stonecroft Boulevard
Chantilly, VA 20151
(703) 633-8300 x7078 (lab)
(703) 633-8300 x8278 (office)
(703) 449-3400 (fax)
eric.kemp at ngc.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mailman.ucar.edu/pipermail/wrf-users/attachments/20080227/592bd93
c/attachment-0001.html
------------------------------
Message: 2
Date: Wed, 27 Feb 2008 13:57:34 -0800
From: "Gustafson, William I" <william.gustafson at pnl.gov>
Subject: Re: [Wrf-users] Experiences with high-resolution WRF-ARW
To: "Kemp, Eric M." <Eric.Kemp at ngc.com>, <wrf-users at ucar.edu>
Message-ID: <C3EB1CCE.1FAE%william.gustafson at pnl.gov>
Content-Type: text/plain; charset="us-ascii"
Eric,
My experience is that high vertical resolutions severely limit the time
step. For example, when running with 57 layers and the vertical
resolution
set to about 35 m at the surface and dx=18, 6, and 2 km for the 3 nested
domains, I need timesteps of 30, 10, 3.33 sec, respectively. How far
down
the timestep has to go will also be a function of the weather conditions
in
the domain. If a frontal passage occurs, or deep convection, with strong
vertical velocities, a shorter timestep is required.
-Bill
On 2/27/08 1:01 PM, "Kemp, Eric M." <Eric.Kemp at ngc.com> wrote:
>
>
> Dear wrf-users:
>
> My group is running WRF-ARW at high resolution (dx = 1 km) to simulate
> turbulence.
> Recently we tried running at very high vertical resolution (dz ~= 12.5
m in
> the lowest
> 500 m of the domain), and encountered CFL errors with time steps as
> low as 1/8 of a second. These runs used the 6th order diffusion
> option from Knievel et al (2007), along with real terrain (area around
> Albuquerque, NM) and physics (MYJ TKE, Noah land surface,
> RRTM/Dudhia radiation, and WSM5 microphysics).
>
> Has anyone here run WRF at such fine resolution, and if so, what were
> your experiences with the time step? The ARW documentation recommends
> 6 seconds for each kilometer of dx, ignoring the effect of vertical
> resolution.
> Are there other settings that should be tuned?
>
> -Eric
>
> Eric M. Kemp
> Meteorologist
> Northrop Grumman Information Technology
> Intelligence Group (TASC)
> 4801 Stonecroft Boulevard
> Chantilly, VA 20151
> (703) 633-8300 x7078 (lab)
> (703) 633-8300 x8278 (office)
> (703) 449-3400 (fax)
> eric.kemp at ngc.com
>
>
>
> _______________________________________________
> Wrf-users mailing list
> Wrf-users at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/wrf-users
--------------------------------------------------------------------
William I. Gustafson Jr.
Atmospheric Science and Global Change Division
Pacific Northwest National Laboratory
3200 Q Ave., MSIN K9-30
Richland, WA 99352
(509)372-6110
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mailman.ucar.edu/pipermail/wrf-users/attachments/20080227/abd7d8b
1/attachment-0001.html
------------------------------
Message: 3
Date: Wed, 27 Feb 2008 13:59:53 -0800
From: David Ovens <ovens at atmos.washington.edu>
Subject: Re: [Wrf-users] Experiences with high-resolution WRF-ARW
To: "Kemp, Eric M." <Eric.Kemp at ngc.com>
Cc: wrf-users at ucar.edu
Message-ID: <20080227215953.GA13350 at frosty.atmos.washington.edu>
Content-Type: text/plain; charset=us-ascii
Eric,
I am sure others can state this more precisely, but I would think your
problem is your use of "real terrain". That sounds like you have not
smoothed the terrain. You still need to remove 2*dx waves from your
terrain or risk CFL errors.
David
--
David Ovens e-mail: ovens at atmos.washington.edu
Research Meteorologist phone: (206) 685-8108
Dept of Atm. Sciences plan: Real-time MM5 forecasting for the
Box 351640 Pacific Northwest
University of Washington http://www.atmos.washington.edu/mm5rt
Seattle, WA 98195 Weather Graphics and Loops
http://www.atmos.washington.edu/~ovens/loops
On Wed, Feb 27, 2008 at 04:01:49PM -0500, Kemp, Eric M. wrote:
>
> Dear wrf-users:
>
> My group is running WRF-ARW at high resolution (dx = 1 km) to simulate
turbulence.
> Recently we tried running at very high vertical resolution (dz ~= 12.5
m in the lowest
> 500 m of the domain), and encountered CFL errors with time steps as
> low as 1/8 of a second. These runs used the 6th order diffusion
> option from Knievel et al (2007), along with real terrain (area around
> Albuquerque, NM) and physics (MYJ TKE, Noah land surface,
> RRTM/Dudhia radiation, and WSM5 microphysics).
>
> Has anyone here run WRF at such fine resolution, and if so, what were
> your experiences with the time step? The ARW documentation recommends
> 6 seconds for each kilometer of dx, ignoring the effect of vertical
resolution.
> Are there other settings that should be tuned?
>
> -Eric
>
> Eric M. Kemp
> Meteorologist
> Northrop Grumman Information Technology
> Intelligence Group (TASC)
> 4801 Stonecroft Boulevard
> Chantilly, VA 20151
> (703) 633-8300 x7078 (lab)
> (703) 633-8300 x8278 (office)
> (703) 449-3400 (fax)
> eric.kemp at ngc.com
>
> _______________________________________________
> Wrf-users mailing list
> Wrf-users at ucar.edu
> http://mailman.ucar.edu/mailman/listinfo/wrf-users
------------------------------
Message: 4
Date: Thu, 28 Feb 2008 10:15:18 -0500
From: "Kemp, Eric M." <Eric.Kemp at ngc.com>
Subject: FW: [Wrf-users] Experiences with high-resolution WRF-ARW
To: <wrf-users at ucar.edu>
Message-ID: <52B67E1F32707847B44B4B7B1238E363020D5690 at xmbv3801>
Content-Type: text/plain; charset="iso-8859-1"
I'd like to thank everyone who responded to my e-mail. There were a lot
of
suggestions made, and I thought it best to respond all at once.
To Nicole Molders: I am using the WSM5 microphysics scheme with this
domain.
I will check to see if heavy precipitation occurs.
To Sen Chiao: I'm using 137 vertical levels. The first 40 levels are
in the
lowest 500 meters (dz ~ 12.5 meters); the next 12 km contain 73 vertical
levels, with dz varying from ~12.5 meters to ~304 meters using a
hyperbolic
tangent stretching function. The remaining 24 vertical levels fill the
domain up to 50 mb. (These vertical resolutions assume the U.S.
Standard
Atmosphere and a reference terrain height of 1631 meters.)
To Joe Galewsky: I have also run WRF using dz = 50 m in the lowest 2
km,
then dz = 125 m for the next 12 km, then 500 m up to 50 mb. Our time
step
for this was 2 sec. I made some attempts at increasing the horizontal
resolution to 0.5 km but experienced CFL errors.
To Chris Walker: I am using the Mellor-Yamada-Janjic PBL scheme, but
with the ratio Km/Kh changed to use the equation from Kondo et al.
(J. Atmos. Sci., 1978) for stable conditions. The 2D Smagorinsky
scheme is used for horizontal deformation. I originally tried using the
3D TKE scheme (km_opt = 2), but this does not work with the land surface
model in WRF 2.2; we may try again with WRF 3.0. I will investigate the
vertical diffusive CFL number.
To David Ovens: I am using the 30 arcsec global terrain available with
WRF. The GEOGRID.TBL.ARW file is standard, and specifies:
* 1 smooth pass
* the smth-desmth_special smoothing option
* average_gcell(4.0)+four_pt+average_4pt interpolation
The documentation for the smoothing and interpolation options is poor,
so
I need to look at the code to see what exactly is going on.
Thanks again!
-Eric
Eric M. Kemp
Meteorologist
Northrop Grumman Information Technology
Intelligence Group (TASC)
4801 Stonecroft Boulevard
Chantilly, VA 20151
(703) 633-8300 x7078 (lab)
(703) 633-8300 x8278 (office)
(703) 449-3400 (fax)
eric.kemp at ngc.com
-----Original Message-----
From: wrf-users-bounces at ucar.edu on behalf of Kemp, Eric M.
Sent: Wed 2/27/2008 4:01 PM
To: wrf-users at ucar.edu
Subject: [Wrf-users] Experiences with high-resolution WRF-ARW
Dear wrf-users:
My group is running WRF-ARW at high resolution (dx = 1 km) to simulate
turbulence.
Recently we tried running at very high vertical resolution (dz ~= 12.5 m
in the lowest
500 m of the domain), and encountered CFL errors with time steps as
low as 1/8 of a second. These runs used the 6th order diffusion
option from Knievel et al (2007), along with real terrain (area around
Albuquerque, NM) and physics (MYJ TKE, Noah land surface,
RRTM/Dudhia radiation, and WSM5 microphysics).
Has anyone here run WRF at such fine resolution, and if so, what were
your experiences with the time step? The ARW documentation recommends
6 seconds for each kilometer of dx, ignoring the effect of vertical
resolution.
Are there other settings that should be tuned?
-Eric
Eric M. Kemp
Meteorologist
Northrop Grumman Information Technology
Intelligence Group (TASC)
4801 Stonecroft Boulevard
Chantilly, VA 20151
(703) 633-8300 x7078 (lab)
(703) 633-8300 x8278 (office)
(703) 449-3400 (fax)
eric.kemp at ngc.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL:
http://mailman.ucar.edu/pipermail/wrf-users/attachments/20080228/a13138f
5/attachment-0001.html
------------------------------
_______________________________________________
Wrf-users mailing list
Wrf-users at ucar.edu
http://mailman.ucar.edu/mailman/listinfo/wrf-users
End of Wrf-users Digest, Vol 42, Issue 14
*****************************************
More information about the Wrf-users
mailing list