fix tsfc on ice in uncoupled non-fractional grid runs

[junwang-noaa]

It is found that the surface temperature (tsfc) is not updated correctly when there is ice on open water grid points in ccpp master branch. This causes the ice temperature difference between gfs.v16 ipd physics and ccpp physics. HREF also noticed some issue on the points with sea ice. The code change is in:

Jun.Wang@hfe06 physics]$ git diff physics/GFS_surface_composites.F90
diff --git a/physics/GFS_surface_composites.F90 b/physics/GFS_surface_composites.F90
index b3000b0..cc61662 100644
--- a/physics/GFS_surface_composites.F90
+++ b/physics/GFS_surface_composites.F90
@@ -547,6 +547,7 @@ contains
             if (.not. flag_cice(i)) then
               tisfc(i) = tice(i) ! over lake ice (and sea ice when uncoupled)
               zorl(i)  = cice(i) * zorl_ice(i)   + (one - cice(i)) * zorl_wat(i)
+              tsfc(i)  = tsfc_ice(i)
             elseif (wet(i)) then
               if (cice(i) > min_seaice) then ! this was already done for lake ice in sfc_sice
                 txi = cice(i) 

[DeniseWorthen]

@junwang-noaa

I've tested this change in my featture/cpld_restart branch and I am now getting c96mx100 and c192mx050 restart repro for fractional grid. Was that expected?

[junwang-noaa]

Yes. I assume your restart test is for the non-fractional grid run using fractional grid ICs. Please let me know if it is a fractional grid run (with frac_grid=.true.)

…

On Tue, Nov 3, 2020 at 2:24 PM Denise Worthen ***@***.***> wrote: @junwang-noaa <https://github.com/junwang-noaa> I've tested this change in my featture/cpld_restart branch and I am now getting c96mx100 and c192mx050 restart repro for fractional grid. Was that expected? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#515 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AI7D6TO4J3UH6ZYMFEMUMKDSOBKG7ANCNFSM4TIZONSQ> .

[DeniseWorthen]

My test was w/ frac_grid=true. This uses the frac_grid ICs.

It is the first time I got restart repro w/ frac_grid, which made me wonder whether this was expected.

[DeniseWorthen]

My mistake--I wasn't comparing the right two runs. I apologize!