access_AocSep2019 - ACCESS-NRI/accessdev-Trac-archive GitHub Wiki

ACCESS Oversight Committee, 16 September 2019

COSIMA members on AOC – Richard Matear and Andy Hogg nominated as alternate members.

ACCESS

  • Model version releases: CMIP6 ACCESS-CM2 versions have been frozen but should wait till Gadi is operational to release (plan for release approx. Dec 2019). Will call it the "CMIP6 submission version" but note that runs done on Gadi will not be bit-reproducible with runs done on Raijin. A draft comms plan will be developed for the next AOC meeting, including user information on how to use the release version.
  • Grand Challenge Simulations: The NCI Director put out a call for a 'grand challenge' ACCESS simulation, to be run early on the new NCI machine Gadi. Proposal from CLEX is to run an Australia-wide simulation at 400m resolution (convection resolving) for as long as we can for a specific extreme weather case or cases. Domain will be consistent with that of the Bureau's National Analysis System currently under development. Start from BARRA reanalysis output. Resource needs are being scoped. Decision waiting on CLEX Board.
  • A summary description of a version of ACCESS NWP model for high resolution regional simulations has been prepared and provided to CLEX for distribution to interested university researchers.

CMIP6 update

  • ACCESS-CM2: PI CTL run at 420 yrs and ongoing. Four SSP scenario runs now almost complete. Historical ensemble (2 members completed, one more planned) on-going. Plan to do an ensemble of three for SSP4-7.0 and possibly other scenarios after Gadi operational.
  • ACCESS-ESM1.5: Three members of each of the four SSP scenarios done. Emission-driven runs being redone following bug fix.
  • Post-processing is progressing, being set up now for NCI to put ACCESS simulation data on the ESG.

ACCESS Science Day, 13 September

  • 11 talks (25 min each), fewer talks and attendees than for the previous Science Day (short period since previous meeting, number of other recent meetings and clashes noted).
  • A dedicated Science Day on CMIP6 results was discussed. Could be held in early Dec, noting CMIP6 papers are due to be submitted for publication for the IPCC AR6 by the end of December. However, workshop on CMIP6 results to be held as part of AMOS Conference in Feb.
  • AOC members to consult with their respective CMIP6 communities on whether there is sufficient interest in a "CMIP6 Science Day" in early December.

HPC issues

Details on the transition to Gadi at NCI have been released, including timelines – note some timelines are short. The transition of data storage from /short to /stash is being worked on.

Other general matters

  • Andy Hogg spoke on COSIMA (Consortium of Ocean and Sea Ice Modelling in Australia). Supports global model suite ACCESS-OM2 – same code as in ACCESS-CM2, with resolutions 1⁰, 0.25⁰, 0.1⁰ lat/lon. All releases are on git-hub and ready to use. Also provides tools to analyse model output. Currently 28 university projects are using the output. The vision of COSIMA is to road test code and resolutions that may become the ocean/sea ice part of ACCESS-CM in the future. OM2 is also being used for participation in OMIP as part of CMIP6. Lead universities are UTas and UNSW. A new COSIMA linkage proposal being prepared for March 2020 submission.
  • Convective Scale UM Modelling Workshop at NCAR (USAF are hosting) in last week of January 2020 – several Bureau and university attendees expected.

Back to AOC meeting notes

Back to main ACCESS wiki page