Fix for specified ice layout non repro issue #200
Conversation
nikizadehgfdl
commented
Sep 13, 2023
nikizadehgfdl
commented
- Closes issue SIS2 in specified-ice mode in AMIP does not reproduce across ice_layout change #199
- Fix ice_layout non-repro of amip using specified ice mode
- Closes issue NOAA-GFDL#199 - Fix ice_layout non-repro of amip using specified ice mode
|
Approved. |
|
There seem to be regressions in two of the tests:
If you have access, you can see which ones here: https://gitlab.gfdl.noaa.gov/ogrp/SIS2/-/pipelines/20620 |
|
Thanks @nikizadehgfdl This fixed the issue for the AM5 amip experiment. It now reproduces across layouts. |
src/ice_model.F90
Outdated
| @@ -1274,7 +1274,7 @@ subroutine set_fast_ocean_sfc_properties( Atmos_boundary, Ice, IST, Rad, FIA, & | |||
| ! timestep. However, it is safe (if wasteful) to call it more frequently. | |||
| if (Rad%do_sun_angle_for_alb) then | |||
| call set_ocean_albedo_from_astronomy(Ice, G, Time_start, Time_end) | |||
| elseif (coszen_changed) then | |||
| elseif (coszen_changed .or. Ice%sCS%specified_ice) then | |||
There was a problem hiding this comment.
Something about this mod is troublesome. If this is to change anything at all, it must be that coszen_change is false but Ice%albedo(k=1) is changing inside set_ocean_albedo_from_coszen. Yet Ice%albedo(k=1) depends only on coszen and should not change if coszen does not!
One way out of this is if Rad%coszen_nextrad has changed since the previous time this sub was called but is still equal to Rad%coszen_lastrad. One way to test this is to set Rad%coszen_lastrad = Rad%coszen_nextrad before line 1258 above.
|
@marshallward any idea when this is going to be merged? |
|
@uramirez8707 We still need to address the regression raised above. AFAIK it is still happening. I will do this test again when I get some time. |
|
I looked further into this. The I am guessing that it was not properly allocated or associated. @nikizadehgfdl can you look into this and try to fix it? |
|
Maybe you can fix this with an |
|
@nikizadehgfdl I can merge this once the possibility of an unallocated |
- Closes issue NOAA-GFDL#199 - Specified-Ice mode needs albedo update at every timestep because coszen changes every timestep - Fix ice_layout non-repro of amip using specified ice mode - Ensure Ice%sCS is associated before checking Ice%sCS%specified_ice
|
Gaea regression: https://gitlab.gfdl.noaa.gov/ogrp/SIS2/-/pipelines/21183 ✔️ 🟡 |
