Heat content associated with fprec and frunoff should enter the ocean at 0 deg C

[gustavo-marques]

As discussed during a MOM6 dev call, if atmos model does not provide the heat content associated with frozen fluxes (e.g., fprec and frunoff), these should enter the ocean at 0 deg C. Currently, the heat content associated with both fprec and frunoff are computed with respect to the temperature of the layer.

See following comment that @adcroft added in MOM_forcing_type.F90:

! fprec SHOULD enter ocean at 0degC if atmos model does not provide fprec heat content.
! However, we need to adjust netHeat above to reflect the difference between 0decC and SST
! and until we do so fprec is treated like lprec and enters at SST. -AJA

Related to #118 #629 #113

[StephenGriffies]

Is that right? What about temp of fprec = min(SST,0) rather than 0C?

[Hallberg-NOAA]

Upon reflection, I agree with @StephenGriffies 's suggestion that min(SST,0) makes more sense for the ocean's guess at the temperature of frozen precipitation and runoff.

[gustavo-marques]

@StephenGriffies and @Hallberg-NOAA can you please clarify why the temperature of fprec (frozen precipitation) can be < 0C?

[StephenGriffies]

Not sure where this issue sits with @gustavo-marques and @Hallberg-NOAA .

My recommendation for min(SST,0) is based on not wanting frozen precip to sensibly warm the SST for cases where the SST is < 0. Instead, with min(SST,0) there is no sensible change in SST when snow falls on it; only latent changes due to melting the snow.

But I agree, this is not an argument based on fundamentals. Instead, what is fundamental is that the atmosphere or a boundary layer should be providing the ocean with this temperature.

What is NCAR doing with this temperature? Any further insights or tests @gustavo-marques ?

[gustavo-marques]

We have been doing the same as GFDL and using SST. I agree that ideally, the atmosphere should provide this temperature - our atmospheric models also lack this capability.
It would be straightforward to implement @StephenGriffies's recommendation and I would be happy to do so.

[StephenGriffies]

@gustavo-marques we (GFDL folks) are following up with hanging MOM6 issues. Have you looked into implementing the min(SST,0.0) issue as an option? Would be nice to have this feature available for reasons noted in this thread.

[gustavo-marques]

@StephenGriffies, after talking to the developers of other CESM components, we have decided to calculate the enthalpy associated with mass fluxes in the coupler to avoid inconsistencies. We are currently working on implementing this in the CESM coupler and follow-up minor modifications in MOM6 will be required. I am happy to also implement the min(SST,0.0) option for fprec and frunoff when I work on these modifications (sometime within the next month).

[StephenGriffies]

Many thanks @gustavo-marques .

[sanAkel]

Hi @gustavo-marques while you work towards implementing ⬆️ could I please suggest that instead of having 0.0C as in min(SST, 0.0) you set a user specified input arg, e.g, Tfreeze_xx = 0.0C for this purpose.

My suggestion is because Tfreeze for fresh/saltwater near river mouths is a function of salinity (or SSS). So may be someone ... someday would like to do a sensitivity study with a different value of the threshold for min value. Thank you! 🙏

[StephenGriffies]

Good idea @sanAkel . Seems like a simple thing to do.