Have flux_atmocn answers changed?

[anton-seaice]

Hi CMEPS folk

We (ACCESS) are trying to update CMEPS versions, we were at CMEPS1.1.2 and have updated to CMEPS1.1.32

I've found changes in all the aoflux fields in the first timestep (e.g. Faox_lat, Faox_evap, _sens, _taox) etc. (I confirmed incoming ocean Import and atm Import fields are unaffected)

Is that expected?

The diff looks like code has been refactored but I can't see anything that has obviously changed. We use

 ocn_surface_flux_scheme = 0

and I set

 aofluxes_use_shr_wv_sat = .false.

to match old behavior

[billsacks]

The main change I'm aware of was #589 (cmeps1.1.14) but setting aofluxes_use_shr_wv_sat = .false. should revert to old behavior, like you said. That said, I'm not sure that was ever tested to confirm that it actually reverted to bit-for-bit behavior with the old code in that respect. (This is awaiting testing in the context of NorESM, which is using the old - false - behavior: NorESMhub#45.)

It looks like #597 made some small changes to the flux code; I'm not sure whether they changed answers.

As you saw, there was a big refactor (in #602 ), but I believe that was bit-for-bit.

@megandevlan are you aware of any other recent changes to the flux_atmocn answers?

[megandevlan]

I wonder if the changes to qsat are responsible for some of the differences? I believe this made things consistent between CAM and CMEPS, but I don't know how much that impacted answers.

10bceb1
579b433

[anton-seaice]

Thanks for confirming my understanding!

Apologies - this was operator error, I'd used CMEPS 1.1.31 and was missing the configurable option for aofluxes_use_shr_wv_sat

[billsacks]

Good, I'm glad you got this resolved!