FSNT Value?

[DrZoidberg]

Hi everyone! I am confused by the expected value of the FSNT variable in comparison to values given by FUS and FDS. I was testing some of my data and found that FLNT = FUL-FDL at the top of the model as expected. I would expect that similarly FSNT would equal FDS-FUS at the top of the model. However, I have found this to not be the case. Does anyone happen to have a suggestion as to what may be happening or the flaw in my reasoning? Thank you!

 

[brianpm]

I don't think I've done that comparison, so I am just guessing. The first thought that come to mind is that somehow the correction for additional absorption above the model top is included in FDS-FUS; in that case, I supposed you should compare with FSNTOA.

Otherwise, I wonder whether you have confirmed if SOLIN = FDS at the top of the model? That is, do you know if there difference is only in the upwelling flux or not?

 

[DrZoidberg]

Thank you for your reply! I noticed for only some (timestep, latitude, longitude) points FDS-FUS and FSNT are equal at the top of the model, but the vast majority of points give extremely different values. I particularly noticed that some FSNT values are zero while FDS-FUS is still in the hundreds.

SOLIN and FDS are pretty close for some values, but again extremely different for most points. I'm concerned there may be an issue with my shortwave data...

 

[brianpm]

I'm not sure how these fields are supposed to be related, and there are some subtle details that happen at the top. It might be worth diving into the code a little bit to see if you are able to figure out if anything happens at (or "above") the top of the model that would make FDS, FUS, SOLIN, and FSNT (FSNTOA?) be different from what you are expecting. I'm going to be working on some radiation things this week, so if I notice anything, I'll let you know.

 

[DrZoidberg]

I'm not sure how these fields are supposed to be related, and there are some subtle details that happen at the top. It might be worth diving into the code a little bit to see if you are able to figure out if anything happens at (or "above") the top of the model that would make FDS, FUS, SOLIN, and FSNT (FSNTOA?) be different from what you are expecting. I'm going to be working on some radiation things this week, so if I notice anything, I'll let you know.

Thank you, I appreciate it! I will keep an eye out and dig into the code.

 

[DrZoidberg]

I'm not sure how these fields are supposed to be related, and there are some subtle details that happen at the top. It might be worth diving into the code a little bit to see if you are able to figure out if anything happens at (or "above") the top of the model that would make FDS, FUS, SOLIN, and FSNT (FSNTOA?) be different from what you are expecting. I'm going to be working on some radiation things this week, so if I notice anything, I'll let you know.

Hello again! I ended up having to work on something else before circling back to this. However, looking at the Fortran code I found this:

fns (1:Nday,pverp-rrtmg_levs+1:pverp) = swdflx(1:Nday,rrtmg_levs:1:-1) - swuflx(1:Nday,rrtmg_levs:1:-1)
fus (1:Nday,pverp-rrtmg_levs+1:pverp) = swuflx(1:Nday,rrtmg_levs:1:-1)
fds (1:Nday,pverp-rrtmg_levs+1:pverp) = swdflx(1:Nday,rrtmg_levs:1:-1)

fsns(1:Nday) = swdflx(1:Nday,1) - swuflx(1:Nday,1)
fsnt(1:Nday) = swdflx(1:Nday,rrtmg_levs) - swuflx(1:Nday,rrtmg_levs)
fsntoa(1:Nday) = swdflx(1:Nday,rrtmg_levs+1) - swuflx(1:Nday,rrtmg_levs+1)

From this, I really don't see why there would be a discrepancy between fsnt or fsns and fds-fus at the corresponding altitude. Did you happen to notice anything related to this in your radiation work? Thank you!