WRFv4.1.3 Initialized w/HRRR Fails at real.exe

msw9090 · Apr 15, 2020

Hello everyone,

Normally I prefer to troubleshoot WPS or WRF on my own, however, I'm running out of ideas to resolve this issue. My procedure to run WRFv4.1.3 utilizing HRRR (form http://home.chpc.utah.edu/) is as follows:

1.a) Obtain HRRR data using Python via the following script <https://gist.github.com/blaylockbk/78d4dd9c159e8974f288b0a57f68c266?fbclid=IwAR31oIBBe9JcQPu7QKq-fIzp5r9skUSIiUMFLuBo_L8U7Y2zbe_kaXNT_tU>
1.b) Download 'prs' and 'sfc' data.
[Using WPSv4.0]
2) Run geogrid.exe [finishes successfully]
3) Run ungrib.exe using the following Vtable <https://github.com/blaylockbk/Ute-WRF-User-Group/blob/master/Blaylock/WPS_for_HRRR/Vtable.HRRR.bkb>. I noticed other, more recent Vtables (for HRRR) fail to extract 3-d variables (I'm running WRFv4.1.3 for 2017-10-28 so I believe the grib tables aren't for the appropriate date). I ran ungrib.exe for both 'prs' and 'sfc' data then combined their ungribed files into met_em*.nc files via metgrid.exe. I also ran ungrib.exe and metgrid.exe using only 'prs' data. Regardless, both processes lead to the same error during real.exe. [finishes successfully]
4) Run metgrid.exe using this METGRID.TBL <https://github.com/blaylockbk/Ute-WRF-User-Group/blob/master/Blaylock/WPS_for_HRRR/METGRID.TBL.HRRR.bkb>. I read in another thread that changing 'FLAG_SOIL_LEVELS' to '0' in METGRID.TBL worked for their simulation(s). I tried running real.exe using the original METGRID.TBL (unchanged) file along with its suggested alteration (FLAG_SOIL_LEVELS==0)... Again, both attempts led to the same error in real.exe. [finishes successfully]
WRFv4.1.3
5) Run real.exe... When I run real.exe without the following namelist.input options:
sfcp_to_sfcp = .true.,
sjust_height = .false.,
interp_type = 1,
lowest_lev_from_sfc = .false.,
force_sfc_vinterp = 1,
etc,
I receive this error: "not enough info for a p sfc computation"
Unfortunately, with the added namelist.input options listed above, I receive this error: "pal, linear is about as low as we go"

I tried different Vtables, METGRID.TBLs, and namelist.input options to solve this problem. Nothing is working at the moment. I haven't made it past this error (usually fails after 10-12 seconds).

I'd really appreciate any suggestions or help that anyone may provide!

I attached my namelist.input, metgrid.log, ungrib.log, METGRID.TBL, and GEOGRID.TBL file. At the moment, I cannot attach the Vtable or rsl.error.out files (I'll add them later if requested!).

Thanks!!!

davegill · Apr 20, 2020

Folks,

This might not an identical error, but it is similar: the real program is failing when using RAP/HRRR hybrid data.

I pick up this file grib2 file from the NCEP ftp servers:

Code:

ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/rap/prod/rap.20200419/rap.t00z.awp252bgrbf00.grib2

I use the ungrib/Variable_Tables/Vtable.RAP.hybrid.ncep. In this file, the soil moisture and soil temperature fields are listed as:

Code:

144 |   1  |   0  |      | SOILM000 | fraction | Soil Moist   0 cm below ground          |  2  |  0  | 192 |   1 |
 144 | 111  |   5  |      | SOILM005 | fraction | Soil Moist   5 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |  20  |      | SOILM020 | fraction | Soil Moist  20 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |  40  |      | SOILM040 | fraction | Soil Moist  40 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  | 160  |      | SOILM160 | fraction | Soil Moist 160 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  | 300  |      | SOILM300 | fraction | Soil Moist 300 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |   1  |      | SOILM001 | fraction | Soil Moist   1 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |   4  |      | SOILM004 | fraction | Soil Moist   4 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |  10  |      | SOILM010 | fraction | Soil Moist  10 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |  30  |      | SOILM030 | fraction | Soil Moist  30 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  |  60  |      | SOILM060 | fraction | Soil Moist  60 cm below ground          |  2  |  0  | 192 | 106 |
 144 | 111  | 100  |      | SOILM100 | fraction | Soil Moist 100 cm below ground          |  2  |  0  | 192 | 106 |
  85 |   1  |   0  |      | SOILT000 | K        | Soil Temp    0 cm below ground          |  2  |  0  |  2  |   1 |
  85 | 111  |   5  |      | SOILT005 | K        | Soil Temp    5 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |  20  |      | SOILT020 | K        | Soil Temp   20 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |  40  |      | SOILT040 | K        | Soil Temp   40 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  | 160  |      | SOILT160 | K        | Soil Temp  160 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  | 300  |      | SOILT300 | K        | Soil Temp  300 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |   1  |      | SOILT001 | K        | Soil Temp    1 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |   4  |      | SOILT004 | K        | Soil Temp    4 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |  10  |      | SOILT010 | K        | Soil Temp   10 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |  30  |      | SOILT030 | K        | Soil Temp   30 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  |  60  |      | SOILT060 | K        | Soil Temp   60 cm below ground          |  2  |  0  |  2  | 106 |
  85 | 111  | 100  |      | SOILT100 | K        | Soil Temp  100 cm below ground          |  2  |  0  |  2  | 106 |

None of these soil fields exist in this grib file.

Code:

> util/g2print.exe GRIBFILE.AAA | grep " 106 "
 106   0    3   0     105       8       0       0     PRES     2020-04-19_00:00:00   00

I do not have any troubles with the 3d atmospheric fields:

Code:

> ncdump -h met_em.d01.2020-04-19_00:00:00.nc | grep -i num_metgrid
	num_metgrid_levels = 51 ;
	float PRES(Time, num_metgrid_levels, south_north, west_east) ;
	float QNI(Time, num_metgrid_levels, south_north, west_east) ;
	float QNR(Time, num_metgrid_levels, south_north, west_east) ;
	float QG(Time, num_metgrid_levels, south_north, west_east) ;
	float QS(Time, num_metgrid_levels, south_north, west_east) ;
	float QI(Time, num_metgrid_levels, south_north, west_east) ;
	float QR(Time, num_metgrid_levels, south_north, west_east) ;
	float QC(Time, num_metgrid_levels, south_north, west_east) ;
	float PRESSURE(Time, num_metgrid_levels, south_north, west_east) ;
	float GHT(Time, num_metgrid_levels, south_north, west_east) ;
	float RH(Time, num_metgrid_levels, south_north, west_east) ;
	float VV(Time, num_metgrid_levels, south_north_stag, west_east) ;
	float UU(Time, num_metgrid_levels, south_north, west_east_stag) ;
	float SPECHUMD(Time, num_metgrid_levels, south_north, west_east) ;
	float TT(Time, num_metgrid_levels, south_north, west_east) ;
		:NUM_METGRID_SOIL_LEVELS = 0 ;

With ncview, I have looked through the fields in the metgrid file.They seem reasonable.

The metgrid program does indicate that it found no soil data, with this global attribute:

Code:

                :NUM_METGRID_SOIL_LEVELS = 0 ;

The expected error that I get from real is:

Code:

-------------- FATAL CALLED ---------------
FATAL CALLED FROM FILE:  <stdin>  LINE:    2669
Not enough soil temperature data for Noah LSM scheme.
-------------------------------------------

davegill · Apr 20, 2020

Folks,
As a follow-up ...
With the WRF release-v4.2 and the WPS release-v4.2 codes, I was able to run a complete end-to-end simulation with the NCAR RAP/HRRR ftp data. The only modification (from the previous post) was that I used the grid 130 data not the grid 252 data from the NCEP ftp site. The grid 130 data has most of the soil data (top level 0 is missing), but enough to run the simulation.

I used this web site for the RAP/HRRR data:

Code:

ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/rap/prod/rap.20200419/

I downloaded these two files:
rap.t00z.awp130bgrbf00.grib2
rap.t00z.awp130bgrbf06.grib2

msw9090 · Apr 29, 2020

Thank you very much for your reply, davegill!

Per your recommendation, I compiled WRFv4.2. I decided to keep WPSv4.0 since it appears that the Vtable was the most significant change regarding RAP/HRRR. I tried running HRRR using this (among others) .grib2 file (with the path included):
<ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/hrrr/prod/hrrr.20200428/conus/hrrr.t00z.wrfnatf00.grib2>
I used the .RAP.hybrid.ncep Vtable, but I ran into the same problem as you (soil data is missing). Compiling WPS can be tricky using my compilers so I decide to find a different solution.

This procedure worked (wrf.exe finished) for me:
1) Download the data...
1.a) https://gist.github.com/blaylockbk/... [see my first post for the full url; extract 'prs' only]
1.b) obtain .grib2 files from the ftp link below [changing the dates since it's operational data;***GET prs .grib2 FILES ONLY*** -> data at specified isobaric surfaces]
example: ftp://ftpprd.ncep.noaa.gov/pub/data/nccf/com/hrrr/prod/hrrr.20200428/conus/hrrr.t00z.wrfprsf00.grib2
2) Run geogrid.exe normally
3) Run ungrib.exe twice
3.a) Use Vtable.raphrrr for the surface/sub-surface fields.
3.b) Use Vtable.RAP.pressure.ncep for 3-d fields
4) Run metgrid.exe combining these two ungribbed files. I used METGRID.TBL.ARW via WPSv4.0.
5) Run real.exe & wrf.exe...

Ironically, the surface pressure interpolation fields weren't necessary for the simulations that I ran.

WRFv4.1.3 Initialized w/HRRR Fails at real.exe

msw9090

New member

Attachments

davegill

New member

davegill

New member

msw9090

New member

Attachments