Verification on Guam grid

[PerryShafran-NOAA]

Hi, there,

Is there any verification option for a grid over Guam, whether it be a mask or NCEP-defined grid or anything you have? There never was any G104 grid for Guam, so I was curious what the best way to verify a Guam grid would be.

Thanks!

Perry

[JohnHalleyGotway]

Hi @PerryShafran-NOAA, I see you have a question about verifying data over Guam. I took a look at the pre-defined NCEP Grid in Office Note 388. I see that grids 175 (10 km lat/lon) and 199 (2.5 km mercator) both mention Guam.

I ran plot_point_obs to visualize these 2 grids in MET:

plot_point_obs out/pb2nc/sample_pb.nc G175.ps -plot_grid G175 -title "Grid 175, 10km Lat/Lon"
plot_point_obs out/pb2nc/sample_pb.nc G199.ps -plot_grid G199 -title "Grid 199, 2.5km Mercator"

The result is extremely uninteresting... just 2 empty plots. That's because the map data bundled with MET does not include an entry for Guam.

Separately, we've discussed defining map data for plots as shapefiles... which certainly would come in handy here!

Let's switch instead to using these grids in gen_vx_mask. NCEP grid 4 is a global 0.5 degree lat/lon grid. Let's select out the G004 points which are included in these Guam grids by running:

gen_vx_mask G004 G175 G004_G175_mask.nc -type grid
gen_vx_mask G004 G199 G004_G199_mask.nc -type grid

Here's the result of plotting the output with plot_data_plane:

Looks like the spatial extent of G175 is much larger than G199.

So you could use either of those grids to define a verification mask covering Guam.
Also remember that you are not limited to pre-defined grids. If you have data in a GRIB file covering Guam, you can point to that file and MET will extract and apply the grid definition from that file. In addition, you can manually define a grid string yourself to use whatever grid you'd like.

Hope that helps clarify.

Thanks,
John HG

[JohnHalleyGotway]

@PerryShafran-NOAA just wanted to follow up on this discussion. Do you have any more questions about this? If not, please feel free to click the Mark as answer button in one of the previous comments.

Thanks!

[PerryShafran-NOAA]

Yes, sorry, I was on leave and put this aside for a bit.

So if I am verifying a Guam forecast file, I can use that very same forecast file to create a mask, but I'll need to use gen_vx_mask to do so. Is that correct?

Thanks!

Perry

[JohnHalleyGotway]

Perry,

I assume you mean that you want to use your Guam forecast file to define a masking region for some larger grid (e.g. like GFS). If so, then yes, it's pretty easy with gen_vx_mask.

For example, here's a CONUS mask of a GFS grid using sample input files named "gfs.grib2" and "conus.grib1":

gen_vx_mask gfs.grib2 conus.grib1 gfs_conus_mask.nc -type grid -name CONUS
DEBUG 1: Input Grid:		gfs.grib2
DEBUG 1: Mask File:		conus.grib1
DEBUG 2: Parsed Input Grid:	LatLon (720 x 361)
DEBUG 2: Parsed Mask Grid:	Lambert Conformal (185 x 129)
DEBUG 1: Output File:		gfs_conus_mask.nc

And here's the result via ncview:

Note that whatever output variable name you use (i.e. -name CONUS) will end up in the VX_MASK column of the resulting .stat output files. So just name it however you'd like that output column to be defined.

[PerryShafran-NOAA]

Actually, the only thing I need a Guam mask for is verifying a Guam grid (the Guam RTMA to be exact). I think that I can use a Guam forecast file once to create the mask, then save the mask in our mask directory and just use that in future Guam verification.

I think I have what I need now, but if I need further help, I'll ask.

Thanks!

Perry

[JohnHalleyGotway]

@PerryShafran-NOAA, ah OK, this is actually much simpler.
If you just want to use all the grid points in the model domain, you can just set:

mask = {
   grid = [ "FULL" ];
...

The "FULL" grid just means exactly that... use all the input point. But that'll put "FULL" in the VX_MASK output column... and that'll appear downstream in METviewer.

If you'd rather VX_MASK be written as "GUAM", then using a gen_vx_mask file with a variable named "GUAM" would produce the desired result.