Using GLDAS for WRF hydro 5.2.0 and crashes

[Mengye Chen]

Dear all,

       I'm currently trying to run WRF-hydro in South America regions. What were your experiences of using GLDAS as the forcing?

       The UCAR provided regridding script doesn't contain the Nan value (-9999) removal. I added this. However, I kept getting error messages after 110-ish time steps regardless of their frequency is hourly, 3-hourly, or daily. It seems to me that something causes the model unstable:

 emitted longwave <0; skin T may be wrong due to inconsistent

 input of SHDFAC with LAI

          43           3 SHDFAC=  0.0000000E+00 VAI=  0.0000000E+00 TV=

   290.0000     TG=   350.3146    

 LWDN=   351.3517     FIRA=  -20745.16     SNOWH=  0.0000000E+00

 FATAL ERROR: STOP in Noah-MP

         Does anyone have any experience with this? Thank you very much,

Mengye Chen, Ph.D.

Center for Analysis and Prediction of Storms

Hydrometeorology and Remote Sensing Laboratory

University of Oklahoma

[Wahidullah Hussainzada]

Dear Chen,

I am receiving same error using WRF output into the model. Had you overcome the issue? If yes, please let me know the sloutions.

Dear all,

Please let me know if you have any solutions to this problem?

With regards

Wahid

[Juan TufBern]

Dear Dr. Mengye Chen,

I hope this message finds you well.

I came across your post on the WRF-Hydro forum regarding the issue you encountered when using GLDAS forcing data for a simulation in South America, specifically the error related to longwave radiation emission and the SHDFAC and LAI parameters in the Noah-MP scheme:

emitted longwave <0; skin T may be wrong due to inconsistent input of SHDFAC with LAI

I am currently facing a similar issue in a simulation using ERA5-Land forcing data. I was wondering if you were able to find a solution or have any recommendations to address this error. Any advice would be greatly appreciated.

Thank you for your time and attention.

Best regards,

[aubrey]

This error is almost always caused by the input met forcing. Make sure you have plausible values for all met variables across the FULL domain - you should not have any "no data" values in the inputs.

Hope that helps!

Aubrey