How best to incorporate Antarctic runoff + calving with JRA forcing

[aidanheerdegen]

We would like to replace the Antarctic runoff in the JRA55-do forcing with a product that splits basal melt from iceberg calving.

We have the data but I'm wondering about the best way to incorporate it into the model.

As we have on-line runoff regridding it would be desirable to not manually alter the runoff files if we don't have to. RYF isn't such a big issue, but when we go to full IAF runs it is not feasible to alter all the input files.

One possibility: add a new calving field into OASIS to insert the calving flux and add a negative freshwater flux at the basal areas to account for the redistribution of freshwater to iceberg calving. Is there a better solution?

ping @nicjhan @russfiedler @aekiss @AndyHoggANU @adele157

[adele-morrison]

Merino's basal fluxes will not necessarily be colocated with the JRA Antarctic runoff. So if we were to follow this method, it would require subtracting off all the JRA runoff, something like:

• calving file: with Merino distributed iceberg FW and negative JRA Antarctic runoff.
• basal melt file: Merino Antarctic basal melt
• JRA runoff: unchanged.

I think it's probably easiest if we just modify the JRA runoff by applying a mask over the Antarctic region and setting everything there to zero. That would be pretty fast and easy to do for the IAF files too. And it only needs to be done once (not every run).

[aekiss]

Just to keep track: @pspence commented at today's MOM meeting that basal melt needs to be input at depth, not surface. Code changes are required to support this. Paul to investigate.

[aekiss]

Basal melt data is not yet published. So we shouldn't put that forcing file on github yet.

[aekiss]

@pspence perhaps coding for basal melt insertion at depth could follow some of what's done for river_insertion_thickness and calving_insertion_thickness? However, these specify the depth range (from surface down) for runoff - not quite the same as insertion of basal melt at depth, and not spatially variable. I guess we need three 2d fields for basal melt: melt mass flux, plus upper and lower depth range for melt insertion? Are they constant in time in the dataset?

[aekiss]

conversation continues here: https://arccss.slack.com/archives/C86NZBNBV

[aekiss]

If we still intend to do this it would be worth seeing what's done in ACCESS-CM2 (the version that @aidanheerdegen and Peter D are harmonising) - as I understand it this accumulates Antarctic P-E and distributes it across the surface of the southern ocean by via seasonally and spatially dependent patterns to represent calving and basal melt freshwater flux from shelves.

[aidanheerdegen]

Good point