ERROR: extrap_type == 2 not implemented for target_z >= zf(1,nz) with GFS data for init_atmosphere

[ep-98d]

I am using GFS data (from here: NCEP Data Products GFS and GDAS), to run the model. But, I am stuck at the step where I generate initial condition. The error message is like the following.

ERROR: extrap_type == 2 not implemented for target_z >= zf(1,nz)
ERROR: *****************************************************************
ERROR: Error in interpolation of t(k,iCell) for k= 1, iCell= 5
CRITICAL ERROR: *****************************************************************

-----------------------------------------
Total log messages printed:
Output messages = 15545
Warning messages = 0
Error messages = 3
Critical error messages = 1
-----------------------------------------

I did some search to find posts asking for similar issue such as this post. Even with changing the config_ztop value to 28000 (or even lower) did not solve the issue. I also tried changing config_nfglevels value to 34 (as the post says) did not help either. I've attached my namelist.init_atmosphere file and the output file from running init_atmosphere.

I even tried it with the exact GFS file that the tutorial mentions which is 2019 GFS file, but I am running into the same problem.

 

[ep-98d]

Looking at the code (mpas_init_atm_vinterp.F) is telling me that config_extrap_airtemp = 'lapse-rate' is responsible for this error. So, I changed it to 'linear' and it seems to be working (no error message regarding extrap_type). But, I wonder if this would prevent me from using lapse rate extrapolation for vertical levels for other cases.

 

[ep-98d]

Problem solved by correctly rebuilding WPS. Using a compiler other than gfortran caused this issue.

 

[mgduda]

Thanks for following up here, and glad to hear that you've gotten past the original error after re-compiling the WPS.