[WRF-Fire V4.6 | MYNN PBL | TKE Budget Output Issue – MYNN3 fails, MYNN2.6 OK]

[yongli wang]

Dear WRF community,
I am conducting a simulation using WRF-Fire V4.6 and encountered an issue with the output of TKE-related variables by using the MYNN2.6 and MYNN3 PBL schemes. When using either MYNN2.6 or MYNN3, the default variable TKE_PBL in the output files is always zero. Therefore, I added the tke_budget = 1, 1, 1 to namelist.input to enable budget diagnostics, then additional TKE budget-related variables are output, including:
QKE: twice TKE from MYNN; QSHEAR: TKE production – shear; QBUOY: TKE production – buoyancy; QDISS: TKE dissipation; QWT: TKE vertical transport; DTKE: TKE tendency term
However, under this setting, the MYNN2.6 scheme runs done normally, while MYNN3 fails to properly write output to wrfout and crashes after running for several hours.

Best regards,
Yongli

 

[Ming Chen]

Yongli,

Would you please upload your namelist.input for me to take a look?

Also, please clarify how you process data (in the WPS stage) for this case.

Where did you download WRF-FIRE V4.6? What is the error message for the crashed case using MYNN3?

I am not an expert in WRF-FIRE, and this issue seems related to WRF-ARW. I will see whether I can figure out the issue. If not, we may need to seek help from our WRF-FIRE expert.

 

[yongli wang]

Ming，
Thank you for your reply, I guess it's the output variable problem of WRF, look at the rsl.error file, the model duration is 10 hours, there should be 11 times data, but in the WRFout, d01 and d02 have only 1 time, d03 is 0 time, but the file size is increasing by hour, the attachments are namelist and rsl, and also use ncdump to view the screenshot of WRFout.
Again,if no tke_budget = 1, 1, 1, and bl_mynn_output = 1, 1, 1, everything is ok. I am testing the other mynn options in the namelist.
The WPS stage referring two new variables ZSF and NFUEL_CAT compared to the original WPS, the two variables processed by the convert_geotiff(GitHub - openwfm/convert_geotiff: A commandline utility for converting GeoTIFF files for use in WRF), the DEM is from the ASTERGDEM 30 meter, the NFUEL_CAT is from the FROM_GLC of the Tsinghua University.
The WRF-Firev4.6 is downloaded the official GitHub.

 

[Ming Chen]

Please upload your namelist.wps, namelist.input and rsl files that contain error messages for me to take a look.

Note that screenshot is hard to read and not easy to search. We don't recommend uploading screenshots because we would like the posts to be searchable for other users who may have similar issues.

 

[yongli wang]

The attach is the namelist.wps,namelist.input and the rsl.error.0000, which is divided to two part01 and part02.
Thanks.

 

[Ming Chen]

Sorry I couldn't find any helpful information in the rsl file....

Would you please try WRFv4.7-Fire ? The developer has made quite a few changes, which hopefully can help overcome the problem.

If it also doesn't work, then I would suggest you compile WRF in debug mode, i.e., ./configure -D, then rerun this case. With the debug mode, you can find exactly when and where the model crashes first. This may give us some clue what is wrong.