Calving flux via iceberg distribution map

[aekiss]

Do we want to distribute the calving meltwater and heat flux according to iceberg distribution? This was done in ACCESS-CM2, but not in the ACCESS-OM2 release (although Pedro Colombo (https://github.com/pedrocol/basal_mom5-collaborative-project) and Jingwei Zhang have explored adding it).

Pedro has also investigated distributing basal and iceberg melt at depth.

[adele-morrison]

I think this is a good idea. What product would we use for the distribution pattern?

When you say heat flux, do you mean adding the latent heat like Pedro has been doing? That seems to have been quite problematic right? (even though physically it seems the right thing to do)

What about dynamic icebergs?

[access-hive-bot]

This issue has been mentioned on ACCESS Hive Community Forum. There might be relevant details there:

https://forum.access-hive.org.au/t/cosima-twg-meeting-minutes-2024/1734/17

[MartinDix]

ACCESS-CM2 uses a fixed iceberg melt distribution pattern scaled by the snow accumulation over the ice sheets. Description paper says "Then we distribute this climatology of NH and SH land ice amounts onto the distributional map of the iceberg climatology from the GC3.1 Lagrangian iceberg model (D. Storkey, pers. comm.)." See Fig 1 of https://www.publish.csiro.au/es/Fulltext/ES19040

The distribution pattern looks like that in Fig 7c of https://gmd.copernicus.org/articles/11/3187/2018/, derived using the NEMO ice shelf parameterization and the interactive icebergs scheme.

[anton-seaice]

See details relevant to JRA55-do here

[anton-seaice]

JRA55-do licalvf field doesn't vary in space or time for Antarctica. The only low-hanging fruit here would be to spread this across the ocean surface with fixed remapping weights. The spreading of the distribution wouldn't change over time (even seasonally).

(Global sum, volume per second:)

Unless i've misunderstood something, even doing what is done in CM2 - where the distribution varies according to a climatology, requires code changes.

Edit: note for myself. The greenland data appears to vary in space and time.

2nd note: The Merino climatology only covers Antarctica and appears to include both basal melt and icebergs in the same data field. Where JRA puts the basal melt with friver and iceberg in licalvf. So using the jra river runoff + merino Antarctic runoff requires notable messing around with removing basal melt from Antarctic from friver and combining the greenland runoff from Jra with antarctic runoff from merino

[chrisb13]

Thanks for checking @anton-seaice. As we chatted this morning, I think a fixed redistribution is still an improvement of putting at the coast (currently done). @aekiss originally mentioned "calving meltwater and heat flux", do we have a field for the latter?

The question then is where to get the spatial map from. Perhaps @MartinDix is able to share the data from Figure 1? And @MartinDix, do you intend to do the same thing in CM3? It would be good to be consistent if practical.

[aekiss]

JRA55-do licalvf field doesn't vary in space or time for Antarctica

Yes, it's constant in time but @anton-seaice are you sure it doesn't vary spatially? There's spatial variation (a different constant for each sector) in calving flux (CF) in Depoorter et al. (2013) supplementary table 1, so I expected this to be inherited by JRA55-do.

Iflicalvf does vary spatially, I guess we'd need to spatially average it anyway in order to then distribute it in proportion to fig 1, right?

[anton-seaice]

JRA55-do licalvf field doesn't vary in space or time for Antarctica

Yes, it's constant in time but @anton-seaice are you sure it doesn't vary spatially? There's spatial variation (a different constant for each sector) in calving flux (CF) in Depoorter et al. (2013) supplementary table 1, so I expected this to be inherited by JRA55-do.

Oh yes, good point - I think what I mean is that the spatial variation doesn't change over time in JRA55do but does in some more sophisticated approaches (e.g. CM2, and Merino).

Iflicalvf does vary spatially, I guess we'd need to spatially average it anyway in order to then distribute it in proportion to fig 1, right?

I think figure 1 is an example but each month has a different spatial distribution?

But to answer your question, the simplistic approach is to sum it all into one bucket and then spread it, but there is possibly a more efficient approach to spreading.

[aekiss]

Is there any reason we can't vary the distribution each month even if licalvf is constant in time? The distribution is just a scaling function (ie we'd need to divide the distribution, eg Fig 1, by its spatial integral so when we multiply by licavf we get the same water flux).

[anton-seaice]

We can do anything ;-)

Um without code changes, there is no provision for remapping weights to change over time.

With code changes or making a new copy of the source data we can vary the distribution over time.

[aekiss]

Um without code changes, there is no provision for remapping weights to change over time.

oh right, sorry I missed that point above. There's a fair bit of temporal variability in Merino et al 2016 supp 2. @MartinDix how strong is the temporal variability in the iceberg distribution you plan to use for CM3?

[aekiss]

"calving meltwater and heat flux", do we have a field for the latter?

Heat flux could be derived from the latent heat to melt the ice, given the ice melt rate and assuming the ice is at its melting point (ie ignoring sensible heat). Might need to consider the Gade relationship following what Pedro did?

[aekiss]

ACCESS-OM2 ignored both the latent and sensible heat needed to melt solid runoff, so it's not a first-order concern.