Session 6: Breakout room 3: Model Configurations/ development - ACCESS-CM3/ESM3

CMIP7 CMIP7 Workshop
,
1

This topic contains discussions, questions and thoughts for Session 6: Breakout room 3: Model Configurations/ development - ACCESS-CM3/ESM3
Time: 11:15 am - 12:45 pm Martin Dix

2

Thanks - are there webex links to join each of these?

3

Thanks - this is breakout 1c from yesterday’s list, right?

4

Yes, it is

5

From Tilo’s talk:

1 Like
6

AH: naming: CM3 should have OM3 in it

TZ: depends how you distinguish configs - eg changes to non-OM components

7

ME: are we aiming for CM3 for CMIP7 or too risky?

MD: want to reflect what’s happening in the communities - don’t want to lose connection to ocean community modelling

8

SM: timeline for UM upgrade

?: no additional science changes in uM but might be diagnostic changes; GC5 not entirely settled; not likely to make science changes

9

Proposed Definitions of model versions
ACCESS-AM3 - basically GAL9
CABLE-3 - Harmonised GitHub version of CABLE
ACCESS-OM3 - MOM6-CICE6
ACCESS-CM3 - ACCESS-AM3+OM3+CABLE-3

1 Like
10

HR: GC5 hight climate sensitivity to be addressed

SO: radiation balance sensitivity testing discovered incorrect sea ice settings

11

I would advocate any interim versions (e.g. ACCESS-AM3+OM2 or ACCESS-AM2+OM3) could be badged as an ACCESS-CM2.5, etc

1 Like
12

?: UKMO aim to correct this in time for CMIP7

?: do we need to wait for updated code of can we use current versions?

MD: won’t be a problem

TZ: we need to adapt what UKMO use - we’ll always be behind

13

BGF: new dot point: aim - what we mean by CM3

  • why would we consider MOM5 in CM3 when there’s a pathway for MOM6?
  • define model components
  • ACCESS land-ice model? can that fit timelines?
14

?: part of the issue is the NUOPC coupler

MD: this is just technical; enough work has been undertaken to know that this will work

?: coupler choice is irrelevant for land surface model (CABLE coupled via JULES)

15

?: more staff available via ACCESS-NRI

16

?: not just a technical issue - need to think through the science as well, eg. JULES science components - eg land issue led to excess sea salt aerosol

  • Jules 7.1 has been moving faster than expected
17

?: timeline for 0.25°

AM: probably this year without WW3; next year with WW3

TZ: COSIMA timeline ends 2025 - could be tight to then put this into a climate model

?: codebase vs configuration - can do a lot of work on early versions

MD: it will be several months before UM-NUOPCis ready

SO: still issues with C-grid coupling - mainly in tripolar region

AH: global low-res has not been top priority to date due to regional; but could priorities global if this is important to have working for CM3

18

ME: is constraint on CM3 due to CMIP7 - just want CMIP7 capability or more/other things?

19

TZ: CMIP7 sets the timeline irrespective of our

MD: CMIP7 timeline not finalised - shortest due date would be hard to reach with NUOPC-based CM

?: beyond CMIP7 - LFRic on horizon - complete rewrite

  • how much do we want capability using absolute latest components

HR: previously have had a “safe” path and an “aspirational” model - should we do this for CMIP7? Gives a fallback if CM3 is too ambitious to reach deadline

SM: expect issues with CABLE coupling; not clear yet re. WOMBAT; also ice sheet coupling

20

AH: define ACCESS-AM3, CABLE3, etc; CM3 shoudl include all these latest components; models that use older components would be CM2.x