access_AocMar2020 - ACCESS-NRI/accessdev-Trac-archive GitHub Wiki
ACCESS Oversight Committee - Notes from the 16 March 2020
1. ACCESS updates
CSIRO is continuing CMIP6 runs for RFMIP (AMIP). An EOI prepared for ALCG computing resources (NCI).
CLEX has noted an intent to retire ACCESS version 1.0 for support from CLEX for new projects. Questions around capability to nudge ACCESS were raised and there seems to be capability linked with chemistry simulations in CSIRO. COSIMA – 1/10th degree model is going ok on gadi with simulations running at 2.5 years per day. Plans to try to do OMIP runs.
BoM notes the upgraded tropical cyclone model becoming operational (4 km relocatable). ACCESS-S2 seasonal system now runs with a locally developed ocean assimilation so is independent of Met Office initial conditions – hindcasts are back on track after an issue with ozone seasonal cycling was fixed. 38 years planned, 1982 to present. Oscar Alves is the point of contact if interested. The UM 11.4 nesting suite is now on gadi.
NCI noted that the Climate DEVL project will wind up soon. They will be collating impacts statements of the value of the CMIP6 infrastructure to include in a report to ARDC for their stage of funding. ESMValTool – users can install on NCI if they want to use. There is currently no community-wide consensus on which evaluation packages would best suit the community. Can anyone with an interest / need for any specific evaluation package please advise their AOC member.
NCI’s ALCG Grants scheme is a new and separate merit allocation scheme that will now run its course. Currently in the EOI stage. NCI also now running a regular small group forum on code modernisation to prepare for software for future HPC systems and new methods. Includes LFRIC - relevant to the UM for future ACCESS. Talk to Ben Evans
The NESP ESSC Hub is in on-going discussions with NCI and CSIRO around visualisation of ACCESS CMIP6 runs.
2. Summary of UM Partner Board Quarterly Videocon Meeting (11 March 2020)
This meeting replaced the annual in-person Board Meeting that was to be held in Delhi but was cancelled due to anticipated travel risk.
Discussion around a possible new joint science activity in seasonal prediction. Several Partners interested, would be based around GLOSEA-5.
The new Cylc-8 scheduler will be based on Python-3. All Partners committed to continued use of Cylc.
Next-Gen modelling systems – the Global Aquaplanet is working but running slowly
Partnership has settled on JEDI for next-gen data assimilation
There is Met Office interest in joint work on climate sensitivity – same UM atmosphere different ocean models, different climate sensitivity.
The UM Users Workshop planned for June 2020 at this stage remains planned to go ahead and be run as a videoconference if required due to travel restrictions.
A GC workshop is provisionally planned to be held in February 2021. More details in due course.
3. Update on Grand Challenge Simulation and STRESS2020 NCI initiative
There are issues with the 400 m resolution Australia run over steep orography and efforts to resolve this are on going. [Update: the issues were resolved and the run is now underway]
In the meantime, an ensemble of 2.2 km simulations for the case originally chosen for the 400 m simulation will be run.
Also aiming to run an ensemble of 10-day simulations for the Townsville flood case at 2.2 km resolution over all of Australia
All runs need to be completed by the end of March at this stage