Cloud sensitivity to CCN-INP using QNWFA_QNIFA_QNBCA_SIGMA_MONTHLY.dat file

D

dnmz24

New member
Hi all,

I will be using the aerosol-aware Thompson scheme to look at the sensitivity of clouds to initial CCN and INP concentrations. I was wondering if I am supposed to change the content of the QNWFA_QNIFA_QNBCA_SIGMA_MONTHLY.dat file to do that. If not, I would appreciate your insights on how I can use this scheme to change the initial CCN-INP concentrations.

Thank you!
 
Please look at the code "module_mp_thompson.F", find how variables from QNWFA_QNIFA_QNBCA_SIGMA_MONTHLY.dat can affect CCN and INP, and move on to modify relevant variables.
Due to limited human power in NCAR, we cannot provide detailed coding instructions.
 
Thank you for your reply. Is it also possible to directly update the W_WIF and I_WIF variables in the met_em files to perform the sensitivity tests using the Thompson aerosol aware scheme rather than updating the source .dat file? Thanks a lot!
 
Thank you. I am running WRF with Thompson aerosol-aware microphysics (mp_physics = 28), use_aero_icbc = .true., and wif_input_opt = 2. I want to do a sensitivity test with very low ice-friendly aerosol / INP input, so I modified I_WIF_FEB in the met_em files and reduced it uniformly by about 10^6 across the domain.

I verified that the modified met_em files contain the expected reduction, and then reran real.exe. However, when I compare QNIFA in wrfinput against the default case, it is only about 100 times lower, not 10^6 times lower. I also noticed that a previous case where I only scaled IFA around the domain and this new near-zero case produce very similar QNIFA values in wrfinput.

At first I checked module_mp_thompson.F and found lower bounds on NIFA, but then realized those apply during wrf.exe, so they cannot explain why the issue is already present in wrfinput after real.exe. I then looked in module_initialize_real.F and found that I_WIF_* is transferred into temporary arrays and then QNIFA is vertically interpolated with vert_interp, but I could not find an obvious direct floor for QNIFA there, although I did find one for QNWFA.

Is there any known lower bound, clipping, background replacement, or interpolation behavior in real.exe for QNIFA / I_WIF_* that could explain why a 10^6 reduction in I_WIF_FEB in met_em leads to only about a 10^2 reduction in QNIFA in wrfinput?

Thanks a lot!
 
Please print the value of the variables below in dyn_em/module_initialize_real.F:
use_aero_icbc
wif_input_opt
wif_fire_emit
They affect how REAL processes the aerosols. Note that wif_input_opt is a derived option and I expect it should have the value of 1.
 
You must log in or register to reply here.
Back
Top