[Met_help] [rt.rap.ucar.edu #72667] History for Grid-Stat error matching grids

Mon Aug 3 11:24:24 MDT 2015

----------------------------------------------------------------
  Initial Request
----------------------------------------------------------------

Hi,

I'm working with METv4.1 and I'm trying to use the Grid-Stat tool to match NCEP Stage IV precipitation to my WRF model output. I used the copy_gb utility to regrid the Stage IV data to my WRF grid. I receive an error from Grid-Stat that my precip and wrf grids don't match, but it seems that they're only off by a tiny amount (and only for Bx and By). See error below:

DEBUG 1: User Config File: GridStatConfig_APCP
ERROR  :
ERROR  : process_command_line() -> The forecast and observation grids do not match: Projection: Lambert Conformal Nx: 499 Ny: 499 Lat_LL: 34.430 Lon_LL: 100.613 Lon_orient: 97.900 Alpha: 12905.872 Cone: 0.598 Bx: 248.9923 By: 8796.5833 != Projection: Lambert Conformal Nx: 499 Ny: 499 Lat_LL: 34.430 Lon_LL: 100.613 Lon_orient: 97.900 Alpha: 12905.872 Cone: 0.598 Bx: 248.9993 By: 8796.6199
ERROR  :

I suspect that the problem is related to the precision of the copygb utility. The central longitude for my WRF grid is -97.89999, but because of the precision of copygb (thousandths of a degree), I'm forced to use either -97.900 or -97.899. I think this may be what is throwing off the Bx and By slightly?

Do you have any suggestions for how I might be able to proceed with Grid-Stat? Are there any other regridding tools that I could use or maybe a flag I could include with copygb to increase the precision of the arguments?

Any advice would be appreciated.

Thanks!
Tricia

----------------------------------------------------------------
  Complete Ticket History
----------------------------------------------------------------

Subject: Grid-Stat error matching grids
From: John Halley Gotway
Time: Mon Jul 20 11:58:55 2015

Tricia,

Yep, you've hit the nail on the head.  Grids specified via copygb are
limited to 3 decimal places of precision.  And that's the source of
the
problem.

The good news is the version of MET we're currently working on will
include
an automated regridding feature.  So once 5.1 is available, all you'll
need
to do is pass grid_stat the original StageIV file and set "to_grid =
FCST"
in the configuration file.  That will help avoid a lot of copygb
headaches,
like the one you're having now.

The bad news is that feature isn't available yet!

Have you post-processed your WRF output using UPP so as to get GRIB
output
files?  If so, I'd suggest also running your WRF output through copygb
to
regrid to a "common" grid... one that only uses 3 decimal places of
precision for the central longitude.

If not, you could modify the grid_stat source code to change that
ERROR to
a WARNING.  Then you'd need to recompile MET.

Or if you can't/don't want to change the source, you could edit the
grid
definition of your WRF data to use 3 decimal places for the central
longitude.  Either edit your model grid and rerun, or tweak the NetCDF
output files to define the grid slightly differently.

Hope that helps.

Thanks,
John Halley Gotway

------------------------------------------------
Subject: RE: [rt.rap.ucar.edu #72667] Grid-Stat error matching grids
From: Lawston, Patricia Marie.[GSFC INTERNS]
Time: Tue Jul 21 07:06:11 2015

Hi John,

Thanks so much for your quick and helpful response. I used UPP and
copygb on my wrf output to make a common grid like you suggested and
it worked great!

Thanks again!
Tricia
________________________________________
From: John Halley Gotway via RT [met_help at ucar.edu]
Sent: Monday, July 20, 2015 1:58 PM
To: Lawston, Patricia Marie. (GSFC-617.0)[GSFC -  HIGHER EDUCATION]
Subject: Re: [rt.rap.ucar.edu #72667] Grid-Stat error matching grids

Tricia,