fix missing connection of ICE and FRZR in some of the SURF components

[zhaobin74]

This PR partially fixed #1110. There will be additional independent efforts coming

NOTE: FRZR is combined with SNO and ICE to form total frozen precip, as done in catchment to keep consistency. FRZR could be part of liquid precip based on discussion. Should that is the case, another PR is needed to change across all surf components. Right now, FRZR is zero so it really does not matter.

Based on the consensus, FRZR should be treated as liquid, so it is combined with PLS and PCU to form RAIN.

The PR is 0-diff for AMIP and non 0-diff for coupled configurations.

Due to scope change, this PR is now non 0-diff.

---

This PR partly fixes #1110.

This PR is 0-diff for AMIP and non 0-diff for coupled configurations. The fixes are implemented in Saltwater and Catchment. For the latter, FRZR was removed from total solid precip.

A separate Non-0-diff PR will address Lake and Landice component.

[zhaobin74]

This PR is in 0-diff uncoupled category but I don't see it in the list. I can change it if someone could point me to the correct label. Thanks.

[gmao-rreichle]

@zhaobin74 : It's probably best to have just one PR for all of the FRZR changes. Then it's easier to make sure everything is consistent across components. Indeed, Lake also needs updating... We can add the Catchment changes to the branch.

[zhaobin74]

@zhaobin74 : It's probably best to have just one PR for all of the FRZR changes. Then it's easier to make sure everything is consistent across components. Indeed, Lake also needs updating... We can add the Catchment changes to the branch.

Sounds good, @gmao-rreichle. I'll update the title to reflect the scope change.

[gmao-rreichle]

@zhaobin74 : It's probably best to have just one PR for all of the FRZR changes. Then it's easier to make sure everything is consistent across components. Indeed, Lake also needs updating... We can add the Catchment changes to the branch.

Sounds good, @gmao-rreichle. I'll update the title to reflect the scope change.

@zhaobin74 , @lcandre2 : Apologies, I didn't realize that the planned Landice changes are non-0-diff. My assumption was that we only reallocate FRZR, which is identical to 0 in the current and forthcoming model versions (v11, v12). This is to make sure FRZR is in the right place if it is ever filled with something. Now I understand that Landice was a bit further behind and had an error in the old precipitation variables. Given that, I'm wondering if it's not better, after all, to separate the changes into two PRs, non-0-diff for Landice and 0-diff for all other components (assuming Lake is also 0-diff). In any case, I will add the Catchment changes to this branch. I'll defer to you about making the non-0-diff Landice changes here or on a separate PR. Apologies for the misunderstanding and confusion.

[zhaobin74]

@zhaobin74 : It's probably best to have just one PR for all of the FRZR changes. Then it's easier to make sure everything is consistent across components. Indeed, Lake also needs updating... We can add the Catchment changes to the branch.

Sounds good, @gmao-rreichle. I'll update the title to reflect the scope change.

@zhaobin74 , @lcandre2 : Apologies, I didn't realize that the planned Landice changes are non-0-diff. My assumption was that we only reallocate FRZR, which is identical to 0 in the current and forthcoming model versions (v11, v12). This is to make sure FRZR is in the right place if it is ever filled with something. Now I understand that Landice was a bit further behind and had an error in the old precipitation variables. Given that, I'm wondering if it's not better, after all, to separate the changes into two PRs, non-0-diff for Landice and 0-diff for all other components (assuming Lake is also 0-diff). In any case, I will add the Catchment changes to this branch. I'll defer to you about making the non-0-diff Landice changes here or on a separate PR. Apologies for the misunderstanding and confusion.

@gmao-rreichle, no problem. You are welcome to add catchment update here. @lcandre2, I am ok with either a non-0-diff PR specific to LANDICE or adding it in this PR.

[gmao-rreichle]

@zhaobin74, @lcandre2 : I added the changes to Catchment and the precipitation corrections in Surface in this commit: 0ef0caa
Besides adding comments, the changes consist of two key edits:

• Remove FRZR in the calculation of solid precipitation.
• Remove FRZR in the calculation of total precipitation when the sum includes large-scale and convective rainfall (PLS and PCU). After talking with @wmputman, my understanding is that FRZR (if it were nonzero) would be included in the sum PLS+PCU. I hope I got this right.

Going forward, Bill said we may want to replace the PLS and PCU inputs with a new export from Moist GC called "RAIN", which does not include FRZR. However, changing from PLS and PCU to RAIN and FRZR would be a bigger undertaking. Note also that PCU and PLS are part of the standard "land forcing" (lfo) file Collection, which provide surface met forcing data for offline (GEOSldas) simulations. I don't have the bandwidth to address this at the moment.

It looks like Lake GC also uses precipitation, so it also needs fixing. Not sure who's in charge of Lake GC.

Please let me know if the above doesn't sound right. Thanks!

cc: @biljanaorescanin

[zhaobin74]

@zhaobin74, @lcandre2 : I added the changes to Catchment and the precipitation corrections in Surface in this commit: 0ef0caa Besides adding comments, the changes consist of two key edits:

• Remove FRZR in the calculation of solid precipitation.
• Remove FRZR in the calculation of total precipitation when the sum includes large-scale and convective rainfall (PLS and PCU). After talking with @wmputman, my understanding is that FRZR (if it were nonzero) would be included in the sum PLS+PCU. I hope I got this right.

@gmao-rreichle, sorry, I am a little bit confused now about FRZR. I get it that FRZR is part of liquid precip. So could you and @wmputman confirm the following?

in current code: total liquid precip = PLS + PCU because FRZR(0 or not) is already included in PCU+PLS

in future: total liquid precip = RAIN + FRZR with RAIN an export from MOIST which does not include freezing rain.

Thanks,

Going forward, Bill said we may want to replace the PLS and PCU inputs with a new export from Moist GC called "RAIN", which does not include FRZR. However, changing from PLS and PCU to RAIN and FRZR would be a bigger undertaking. Note also that PCU and PLS are part of the standard "land forcing" (lfo) file Collection, which provide surface met forcing data for offline (GEOSldas) simulations. I don't have the bandwidth to address this at the moment.

It looks like Lake GC also uses precipitation, so it also needs fixing. Not sure who's in charge of Lake GC.

@lcandre2 and I will work on Lake component.

Please let me know if the above doesn't sound right. Thanks!

cc: @biljanaorescanin

[zhaobin74]

Thank you, @lcandre2, for fixing the Lake GC. This PR is ready for review.

[lcandre2]

Actually, I just pushed two small changes. Now it is ready.

[gmao-rreichle]

@zhaobin74 @lcandre2 : Is this PR still 0-diff for AMIP (as indicated by the label) after the Lake changes? Wouldn't the (previously missing) icefall change the surface temp at least a little bit?

[lcandre2]

@zhaobin74 @lcandre2 : Is this PR still 0-diff for AMIP (as indicated by the label) after the Lake changes? Wouldn't the (previously missing) icefall change the surface temp at least a little bit?

Great question and I was pondering tha myself last night. My guess is it no longer zero diff, but I haven't run the tests -- happy to, but the exact tests weren't documented on the wiki.

[gmao-rreichle]

@zhaobin74 @lcandre2 : Is this PR still 0-diff for AMIP (as indicated by the label) after the Lake changes? Wouldn't the (previously missing) icefall change the surface temp at least a little bit?

Great question and I was pondering tha myself last night. My guess is it no longer zero diff, but I haven't run the tests -- happy to, but the exact tests weren't documented on the wiki.

FYI, tests are failing right now because I updated the branch after #1112 was merged into develop (reflexively, without thinking...). Now we need to wait for GEOSgcm to catch up (GEOS-ESM/GEOSgcm#938). Sorry for the confusion

[zhaobin74]

@zhaobin74 @lcandre2 : Is this PR still 0-diff for AMIP (as indicated by the label) after the Lake changes? Wouldn't the (previously missing) icefall change the surface temp at least a little bit?

Great question and I was pondering tha myself last night. My guess is it no longer zero diff, but I haven't run the tests -- happy to, but the exact tests weren't documented on the wiki.

@lcandre2 and @gmao-rreichle, you are both right. With the Lake change, it is no longer 0-diff. Maybe we can just bring all together into this PR. I'll update the label. Sorry, I missed that.

[github-actions]

This PR is being prevented from merging because you have not added one of our required labels: 0 diff, 0 diff trivial, Non 0-diff, 0 diff structural, 0-diff trivial, Not 0-diff, 0-diff, automatic, 0-diff uncoupled, github_actions. Please add one so that the PR can be merged.

[sdrabenh]

@lcandre2 this PR has been merged into develop. Please make a new PR into develop once you made the companion lake and landice changes.