./create_forcing.exe namelist.input.ERA5 ERROR!

[prolay0101]

Dear Sir,
I encountered an error while executing ./create_forcing.exe namelist.input.ERA5:

############################
 You have this many undefined T points over valid land points:            1
 This tends to happen around coastlines or for unresolved islands in your forcing data
 ===== Consider increasing expand_loop in namelist =====
 However, this may also occur if your forcing source does not cover your domain.
###########################

This issue occurs when I run the model over a larger Indian domain (geo_em.d01.nc). However, the smaller nested domain (geo_em.d02.nc) runs without any problems.
I have already tested multiple values for EXPAND_LOOP (10, 20, 25, 50) and also set MASK_WATER = .FALSE., but the error persists.

I have attached the relevant data and namelist files in a zip archive for reference.

https://drive.google.com/file/d/1Kz7vPyAYhvhjdu7X7Q5T6TsafSDrDL-D/view?usp=drive_link

[prolay0101]

Dear Sir,

After several trial-and-error attempts to determine suitable parameter settings for successfully running "create_forcing.exe", I found that the following configuration works well for my domain geo_em.d01.nc:

"EXPAND_LOOP = 75"
"MASK_WATER = .FALSE." (also tested with ".TRUE.")

For the nested domain geo_em.d02.nc, even a smaller value such as "EXPAND_LOOP = 3" appears to work.

As I intend to use the generated output from HRLDAS as forcing input for WRF, I would be grateful if you could kindly clarify a few points:

1. Is it acceptable to use different integer values of "EXPAND_LOOP" for the nested domains (i.e., "geo_em.d01.nc" and "geo_em.d02.nc")?
2. Whether setting "MASK_WATER = .FALSE." or ".TRUE." would have any impact on subsequent simulations in HRLDAS and WRF?
3. Since I have compiled using "mpif90", is it appropriate to execute "create_forcing.exe" in parallel using a command such as:
   "mpirun -np 32 --bind-to core --map-by core ./create_forcing.exe namelist.input.ERA5"?

[cenlinhe]

1. Yes, it is fine to use different values for "EXPAND_LOOP" for the nested domains.
2. MASK_WATER = .FALSE. or .TRUE. affects your forcing and initial condition for HRLDAS run only. suggest using mask_water = .FALSE.
3. create_forcing.exe cannot use parallel processing. To speed up, you can create multiple create_forcing run folders and manually launch the create_forcing.exe in each individual folder together to process forcing for different time slices of your study period.

[prolay0101]

Dear Sir,

Thank you for the kind explaination!