[Wrf-users] Bad TSOIL values in GFS 0.25° GRIB2 data

[HEDDE Thierry 137147]

Hi to all WRF users,

Apparently, there was a change in the GFS 0.25° DATA on July 19th 2017 from http://nomads.ncep.noaa.gov/ 
Since this date the TSOIL data show some 0K values over land which lead to a quick crash in WRF.
Does somebody have or had the same problem? I've read there was some weird stuff about soil temperature in GFS (see Vtable file).
I can't find any other GFS variable containing Soil Temperature value. The TMP variable doesn't have a soil level.
If anybody solved this and could share it I would appreciate! 

Details from my inquiry:
---------------------------------
I was using WPS/WRF 3.5.1 as an operational daily but since this problem occurred I've switched to 3.8.1 in case the problem would disappear, but it didn't.
WRF ends with segmentation fault, and a warning from NOAH LSM "Flerchinger USEd in NEW version. Iterations=          10".
I'm using the classical 3.8.1 Vtable for GFS with the following table for Soil T:
  11 | 112  |   0  |  10  | ST000010 | K       | T 0-10 cm below ground layer (Upper)    |  0  |  0  |  0  | 106 |
  11 | 112  |  10  |  40  | ST010040 | K       | T 10-40 cm below ground layer (Upper)   |  0  |  0  |  0  | 106 |
  11 | 112  |  40  | 100  | ST040100 | K       | T 40-100 cm below ground layer (Upper)  |  0  |  0  |  0  | 106 |
  11 | 112  | 100  | 200  | ST100200 | K       | T 100-200 cm below ground layer (Bottom)|  0  |  0  |  0  | 106 |
  85 | 112  |   0  |  10  | ST000010 | K       | T 0-10 cm below ground layer (Upper)    |  2  |  0  |  2  | 106 |
  85 | 112  |  10  |  40  | ST010040 | K       | T 10-40 cm below ground layer (Upper)   |  2  |  0  |  2  | 106 |
  85 | 112  |  40  | 100  | ST040100 | K       | T 40-100 cm below ground layer (Upper)  |  2  |  0  |  2  | 106 |
  85 | 112  | 100  | 200  | ST100200 | K       | T 100-200 cm below ground layer (Bottom)|  2  |  0  |  2  | 106 |
  11 | 112  |  10  | 200  | ST010200 | K       | T 10-200 cm below ground layer (Bottom) |  0  |  0  |  0  | 106 |

I'm working over France/Europe, and as an example, I've noticed O K values for ST000010 from met_d files over Corsica.
It's always at the same places the 0 K soil T is located. Here for example around 8.8° E, 42.4°N but it's not the only place as I found another place over Italy.
Please find the Panoply image here: ftp://ftp.cea.fr/incoming/y2k01/ST000010 in met_em.d01.2017-08-27_12_00_00_corsica.png

Extracting directly the TSOIL values from GFS files using "g2print.exe -V" confirms that it's the GFS TSOIL field which contains 0 values.

Thanks for your answers!
 
Thierry HEDDE   

Laboratoire de Modélisation des Transferts dans l'Environnement
CEA/Cadarache
SIRET : 775 685 019 00587
Bât 727 P102
13108 Saint Paul lez Durance cedex
FRANCE
 43.6877°N  5.7611°E     
   





-------------- next part --------------
A non-text attachment was scrubbed...
Name: ST000010 in met_em.d01.2017-08-27_12_00_00_corsica.png
Type: image/png
Size: 29394 bytes
Desc: ST000010 in met_em.d01.2017-08-27_12_00_00_corsica.png
URL: <http://mailman.ucar.edu/pipermail/wrf-users/attachments/20170828/71c0d485/attachment-0001.png>