How to get diagnostic radiation flux variables with RRTM?

[yhanw]

Hello Sir/Madam,

I'm running WRF-Chem with RRTM radiation schemes (ra_lw_physics = 1; ra_sw_physics = 1). The run is successful, however, there are very limited diagnostic flux variables in outputs (e.g., SW/LW flux at TOA, clear-sky, etc.). I tried using runtime I/O field to add variables like TLWDN, SLWDN etc., but it does not work, because these state variables are not part of rrtmscheme/swradscheme package in Registry.EM_COMMON.

I'm hoping to ask if there is a way to turn on these variables while sticking to RRTM? Or does that mean I must switch to other schemes like RRTMG for accessing diagnostic variables? I'd appreciate any advice on this issue.

Thank you in advance for your time!

Best,
Yuhan

####################### Registry.EM_COMMON #############################

package rrtmscheme ra_lw_physics==1 - -

package rrtmg_lwscheme ra_lw_physics==4 - ozmixm:mth01,mth02,mth03,mth04,mth05,mth06,mth07,mth08,mth09,mth10,mth11,mth12;state:aclwupt,aclwuptc,aclwdnt,aclwdntc,aclwupb,aclwupbc,aclwdnb,aclwdnbc,lwupt,lwuptc,lwdnt,lwdntc,lwupb,lwupbc,lwdnb,lwdnbc,o3rad

package swradscheme ra_sw_physics==1 - -

package rrtmg_swscheme ra_sw_physics==4 - ozmixm:mth01,mth02,mth03,mth04,mth05,mt
h06,mth07,mth08,mth09,mth10,mth11,mth12;state:acswupt,acswuptc,acswdnt,acswdntc,acswupb,acswupbc,acswdnb,a
cswdnbc,swupt,swuptc,swdnt,swdntc,swupb,swupbc,swdnb,swdnbc,o3rad
####################################################

 

[kwerner]

Hi,
Unfortunately to get those specific variables (TLWDN and SLWDN), you would need to use the Goddard radiation scheme (option 5). The alternative would be to make modifications to the option 1 schemes to include those variables, but that would probably be a lot more work.

 

[yhanw]

Hello Kelly,
Many thanks for your response! I tried adding state variables to option 1 in registry - they do exist in wrfout now but just show 0 values everywhere. I'm assuming source code changes will be needed to really do the calculation, and like you said, it'll be a lot more work. So I am planning to switch to RRTMG (option 4) or Goddard (option 5).
I do hope to ask: would it be possible to use my old wrfrst file (with RRTM) to continue the run with RRTMG/Goddard? I'm unsure if I should rerun everything from the start using RRTMG/Goddard, which could take lots of CPU/time. I'd also be grateful to learn if you have any recommendations for choosing between RRTMG versus Goddard. I'm simulating heatwaves in India.
Thank you!
Best,
Yuhan

 

[kwerner]

Apologies for the delay in response. I believe you should be able to change the radiation between the original run and a restart. I ran a short test to make sure I didn't receive any errors related to making that change, and I didn't, so it should be okay. Just check your output before it gets too far into the simulation to make sure everything looks as is expected.

Unfortunately I don't have experience that compares RRTMG/Goddard. I would suggest running for a short period of time with each option and see which is producing the most realistic outcome.

 

[yhanw]

Apologies for the delay in response. I believe you should be able to change the radiation between the original run and a restart. I ran a short test to make sure I didn't receive any errors related to making that change, and I didn't, so it should be okay. Just check your output before it gets too far into the simulation to make sure everything looks as is expected.

Unfortunately I don't have experience that compares RRTMG/Goddard. I would suggest running for a short period of time with each option and see which is producing the most realistic outcome.

Thank you Kelly!! This is super helpful.