First day output of FSNS in a hybrid run

[kezhoulumelody]

Hi,

I found a weird looking FSNS pattern in the first day of any type of hybrid run. I guess this is part of the model spin-up process as the second day looks completely normal. Could anybody explain why the first-day net short wave radiation pattern looks particularly like this?

Thank you so much!

Melody

For reference, the case displaying here has a compset B1850, reference case b.e21.B1850.f09_g17.CMIP6-piControl.001_v2 at 0501-01-01

 

[sacks]

I think what you're looking at is the 0th time step: there is an oddity in CAM and CLM output where the first history file written is from a single initialization time step rather than a full averaging period. Would that help explain what you're seeing? See Improve handling of 0th / initial time in history output · Issue #459 · ESCOMP/CAM for some more information.

 

[kezhoulumelody]

I think what you're looking at is the 0th time step: there is an oddity in CAM and CLM output where the first history file written is from a single initialization time step rather than a full averaging period. Would that help explain what you're seeing? See Improve handling of 0th / initial time in history output · Issue #459 · ESCOMP/CAM for some more information.

Hi Bill, that helps a lot. Is there a way to avoid this oddity? I want my simulation to hybrid from a certain initial land surface condition, and I want to look at the instant changes. Does that mean I should discard the first history file? Thanks

 

[sacks]

Yes, my understanding is that the right thing to do is to just discard the first history file - at least if it has a time stamp of 0001-01-01-00000, like this one does. (Monthly history files should be fine: I don't think those have this extra 0th time step issue.)

 

[kezhoulumelody]

Yes monthly mean files do not have this issue. Thanks, Melody