Init_Atmosphere, ERA5, CRITICAL ERROR: Error in interpolation of st_fg to MPAS grid: num_st = xxx

[yibssss]

Hi Yall!
I encountered
CRITICAL ERROR: Error in interpolation of st_fg to MPAS grid: num_st = 2168
issue while running the init_atmosphere for a limited area MPAS run over CONUS, using ERA 5 dataset.
I read some of the other post and tried switching my domain but it always result in
CRITICAL ERROR: Error in interpolation of st_fg to MPAS grid: num_st = xxxx, where xxx is some number (i'm assuming the grid number?)
The same ERA5 input condition has ran WRF 4.5 without issue

 

[Ming Chen]

This is more like a data issue. We are aware that incorrect processing of ERA5 may lead to such errors.

Would you please run the script to process ERA5, then rerun init_atmosphere using the newly created intermediate files?

Please let me know if you still have any problems. Thanks.

 

[mgduda]

Aside from a potential issue with the soil temperature data in the intermediate file, it also appears from the log.init_atmosphere.0000.out file that the ERA53D intermediate file may be lacking any 3-d humidity information -- either relative humidity or specific humidity. At isobaric levels, only GHT, TT, U, and V are reported; e.g.,

Interpolating GHT at 4 95000.0
Interpolating TT at 4 95000.0
Interpolating U at 4 95000.0
Interpolating V at 4 95000.0

 

[yibssss]

Thanks for the help! I didnt get to use the ERA5_TO_INT script since I havnt tried downloading NC form of ERA 5 from RDA or on glade computer.
I included both rel humidity and specific humidity + increased my crop of ERA5 data and it seems working now!
Only issue is i have to change the vertical interp to "Linear" otherwise this happens
----------------------------------------------------------------------
Beginning MPAS-init_atmosphere Error Log File for task 25 of 32
Opened at 2025/10/16 13:39:31
----------------------------------------------------------------------

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= 1
CRITICAL ERROR: *****************************************************************
Logging complete. Closing file at 2025/10/16 13:39:31

Thanks Admins!

 

[mgduda]

The error messages you were seeing related to vertical interplation / extrapolation,

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= 1
CRITICAL ERROR: *****************************************************************

generally turn up when your model top is above the highest level in the first-guess date (ERA5 in this case). From your log file, it looks like you have ERA5 levels going up to 1 hPa. Vertically extrapolating by a small distance is usually not problematic, but it might be worth considering the ERA5 model-level dataset if your model top (config_ztop in the &vertical_grid namelist group) is significantly (i.e., more than a couple of km) above 1 hPa.