Vertical turbulent diffusion probability lower bound causing stalled simulations

[samfrederick]

I noticed the repository version of the diffusion code still has the lower probability bound on vertical turbulent diffusion in interface/wrf_pmc_trans_aero.F90. This was leading to hanging simulations and appears to occur earlier as we increase the number of particles. Adding the following commented lines (starting around line 1510) seems to fix the issue:

do k = pmc_ks,pmc_ke if (k .ne. k_index) then prob = env_state%prob_vert_diffusion(k,i_class) ! if (prob > 1.0d-6) then dest_vol = env_states(i_index,k,j_index)%cell_volume flux = prob * num_conc_old num_conc_new(i_index,k,j_index) = & num_conc_new(i_index,k,j_index) + flux * & (dest_vol / source_vol) num_conc_new(i_index,k_index,j_index) = & num_conc_new(i_index,k_index,j_index) - flux ! end if end if end do