Change range of solid ice runoff removal from 60S to 57S

[darincomeau]

Extends the northern range of solid ice runoff from Antarctica from 60S to 45S 57S when the ocean's namelist option config_remove_ais_ice_runoff = .true.. The new solid ice runoff mapping with increased smoothing for the SOwISC12to30E3r3 mesh in #6759 has a small amount of runoff north of 60S, which needs to be removed when Antarctic solid ice runoff is intended to be removed.

[BFB]
All current configurations that do not use the mapping file in the above PR do not have solid ice runoff between 60S and 45S 57S, so there is no impact.

[darincomeau]

Thanks to @cbegeman for noticing the circled sliver of AIS runoff from the increased smoothing mapping file that is being removed here:

I'm including this in new BlueTip test run, will post results here after a decade showing it is removed.

[xylar]

Do we ever have ice runoff in places like Patagonia?

[darincomeau]

Do we ever have ice runoff in places like Patagonia?

I was worried/wondering about that too - it doesn't appear in this plot, but should confirm.

[xylar]

@stephenprice, @matthewhoffman, @trhille, @chloewhicker, could any of you answer the question above? We're wanting to know if it's safe to remove ice runoff further north or if we will ever start to remove ice from Patagonia in any E3SM simulations. In other words, do we currently (or will we soon) have configurations with mountain glaciers that produce ice runoff?

[stephenprice]

Refresh my memory -- Does the iceRunoffFlux field represent calving?

[darincomeau]

Does the iceRunoffFlux field represent calving?

No, snowcapping solid ice runoff. @cbegeman is double-checking this (thanks!).

[stephenprice]

Ok, thanks. I would say that, for now, you do not need to worry about snowcapping outside of Antarctica and/or Greenland. We don't have any snowcapping related treatment of mountain glaciers and/or icecaps right now. These are just treated as areas covered w/ seasonal snow at the moment (this will need revisiting at some point in the future).

@matthewhoffman should confirm if he agrees with me on this or not.

[matthewhoffman]

@chloewhicker , can you confirm what happens over non-ice-sheet locations/land-units if the maximum snow thickness is reached?

[xylar]

I would have thought that ice runoff would also have to include any calving from non-ice-sheet glaciers. I don't think there is any other possible term for this.

[chloewhicker]

I've never seen enough snow over mountain glaciers to cause snow capping. The mass balance (QICE) over them is consistently negative - I think the SMB shown is just artificially "melted down" into the infinite reserve of land ice and never actually sent anywhere because there is no ice dynamics underneath this land ice in ELM and we don't see negative runoff or a large liquid flux into the ocean in these areas (correct?).
Here are ELM plots of the average snow depth (note odd colorbar to show range) and the SMB over valid land ice regions (which includes mountain glaciers).

[xylar]

@darincomeau, I think this will be non-BFB for any DIB test cases, right?

Or is it BFB because no existing tests have such an aggressive spreading function as you showed above?

[cbegeman]

So based on what @chloewhicker said, it seems that we could have snowcapping fluxes from Patagonia, even though we currently don't in configurations so far. I will calculate the maximum latitude in the southern hemisphere that has ice runoff fluxes in the current spreading file so we can choose that latitude as the cutoff for ice runoff removal rather than the more generous 45 degS.

[matthewhoffman]

Thanks, @chloewhicker . I believe the configuration that is being used for the ocean model has the shallow 1-meter snowpack model, and could be using a range of different atmospheres. So we can't rely on the same snow depth and SMB behavior as you showed above to avoid a potential issue here. E.g. it looks like there are some positive SMB locations on Siberian islands and in Alaska. Do you know what happens to non-glc land-unit cells in ELM if the snowcapping depth is reached?

[chloewhicker]

if the snow depth is > than the max snow depth and the new routing method is not being used (freeze precip before sending it to river runoff) then any additional precip will be directly routed to runoff in the state it fell in. Feel free to point me to the ELM and MOSART history fields and I can look at them

[darincomeau]

@darincomeau, I think this will be non-BFB for any DIB test cases, right?
Or is it BFB because no existing tests have such an aggressive spreading function as you showed above?

This would only be non-BFB if config_remove_ais_ice_runoff = .true. and there is solid runoff in test cases between 60S and 45S. The only possibilities are cryo configurations where i) spreading function extends AIS solid runoff north of 60S, or ii) there is non-AIS based solid runoff in that band.

For i), I don't think there are any tested cryo configurations (B- or G-cases) that use an extended spreading function.
For ii), G-cases do not have solid ice runoff apart from Antarctica/Greenland, and I'm gathering from the (very useful) discussion here that it's possible, but not currently happening, for B-cases to have non-ice sheet solid ice runoff to exist in this band.

It sounds like it would be safer to adjust the bound here closer to the problem area, rather than the judicious 45S I suggested.

[xylar]

Okay, great. So we have a reasonable expectation of being BFB.

[matthewhoffman]

Thanks, @chloewhicker . Based on this statement:

then any additional precip will be directly routed to runoff in the state it fell in.
it seems like the configuration in this PR is susceptible to inadvertently zeroing solid runoff in Patagonia if it should ever occur.

@cbegeman and @darincomeau , if you think you can define criteria to avoid non-AIS runoff getting zeroed, that seems easiest. I do wonder with this more extensive spreading function if even if you carve out Patagonia from the mask, it's still possible to inadvertently zero solid runoff from islands in the Southern Ocean - I see at least 3 showing up in Chloe's ELM figures.

[cbegeman]

A bit of a complication: we do get ice runoff all over Patagonia in our latest v3 SORRM alfred1 run. It's just super tiny (like 1e-21 kg m^-2 s^-1 rather than 1e-7). So I think we're probably looking at non-BFB.

The maximum southern hemisphere latitude that is associated with the band shown above is -57.1854.

[darincomeau]

-57 is just south of southern tip of South America, so we should be good with that?

[cbegeman]

Yes, I don't see any issue with using -57. I haven't checked the latitude of those islands @matthewhoffman pointed out, but I think that creating a more complex mask is probably more effort than it's worth.

[darincomeau]

Oops I missed @matthewhoffman 's comment when I wrote that.

[darincomeau]

Ice runoff north of 60S now removed:

[xylar]

This looks reasonable to me.