I tried different versions but always get the same error.
Could you share your runscript that I can compare mine? Or could you take a look into mine? Maybe it is some other setting?
/work/bb1337/icon-mpim_10.2025/run/exp.atm_xpp_R2B4_90lev_1step_fg.run
I took the liberty of copying your ICON checkout. I’ve compiled it with config/dkrz/levante.gcc and I’m getting out-of-bounds errors in the function that parses the group:… items. Looks like the carbon cycle allocates a metric ton of additional variables and overloads the fixed-size arrays for the varlists. The extra items spill over into the pl_varlist, which is directly behind it in the output namelist structure. That explains your error. Could you increase mo_impl_constants:MAX_NVARS and max_var_ml to 2000 and try again?
1 Like
It worked, thank you very much!
I will proceed and come back to you when I stumble over a new error 
Hi, @rwith and @drieger, you already helped a lot! However, I stumble from on error to another.
The initialisation kind of works, the ecrad error seems to be gone. I remapped the fg-file from the R02B04 to R02B05, R02B06 and R02B07. It seems that I get to high pressures due to the mismatch between the orographic elevation between the domains and the wind field that come from the 160km Domain which has much flatter terrain.
Is there a solution for this problem? I assume (hope) that this is not the first time this happens
I have not yet stumbled across such an issue, but this is a general recommendation to avoid too high resolution jumps during interpolation. Sometimes a smaller time step helps to ‘digest’ this initial shock.
Alternatively, you could try a step-by-step approach to go to R2B5, perform a short forecast and use the resulting data to interpolate to R2B6 and so on.