In rrtmg_lw_rtrnmr.f90, subroutine rtrnmr, there are several suspect-looking accesses to cldfrac(lev-1) in the loop over lev ending at lev=1. I'm hitting this one in a DALES run which happens to have cloud in the lowest model layer. A debug build stops with the message
At line 499 of file /home/jansson/dales-tester/dales/external/RRTMG/RRTMG_LW/src/rrtmg_lw_rtrnmr.f90
Fortran runtime error: Index '0' of dimension 1 of array 'cldfrac' below lower bound of 1
It's of course possible we're doing something wrong on the DALES side. But looking at the RRTMG_LW code, it seems if the lowest level is cloudy there is an access out of bounds.
In rrtmg_lw_rtrnmr.f90, subroutine rtrnmr, there are several suspect-looking accesses to
cldfrac(lev-1)in the loop over lev ending at lev=1. I'm hitting this one in a DALES run which happens to have cloud in the lowest model layer. A debug build stops with the messageIt's of course possible we're doing something wrong on the DALES side. But looking at the RRTMG_LW code, it seems if the lowest level is cloudy there is an access out of bounds.