Science1 publisher2 min readPublished
ACCESS-NRI puts Australia's ice-sheet modellers on a single supported ISSM build
Australia's climate simulator now offers one shared ISSM build, a Python interface and a community data pool to its ice-sheet researchers. Coupling that ice to the ocean and atmosphere inside ACCESS comes later.
The Scientist · Science desk
What happened
- ACCESS-NRI has released four components of the Ice Sheet and Sea-level System Model workflow to Australian researchers, calling it the first step toward a dynamic ice-sheet model inside the ACCESS climate suite.
- Until now the suite has used a static proxy to approximate the impact of ice sheets, and Australian work has focused on individual regions instead of the whole of Antarctica, the team lead said.
- One release is a supported ISSM deployment on the NCI Gadi supercomputer, intended to replace locally compiled versions with a single open and reproducible build.
Compiled by The ScientistSomething wrong?How this is made
Why it matters
- constraint The coupled configuration does not exist yet, so any sea-level figure coming out of ACCESS still carries the static proxy's provenance, and downstream users quoting it are quoting the old arrangement.
- capability Two Australian groups can now compare ISSM runs without first reconciling which local compile produced them, because the supported build and the shared data pool are the same for both.
- decision Groups maintaining their own ISSM builds have to decide whether to keep doing so, given that the training materials and workshops are attached to the supported deployment.
Ice sheets are the largest drivers of global sea-level change on centennial and longer timescales. A dynamic model captures the interaction in both directions: ice delivers freshwater to the ocean, changing temperature, salinity and circulation, the reflective ice surface acts on the atmosphere, and changes in ocean and atmosphere then feed back into how the ice evolves.
What has shipped is tooling. Two of the four components are for running the model and two are for getting at data. ACCESS-ISSM is an ACCESS-NRI-supported deployment of the ISSM code on Gadi, the National Computational Infrastructure supercomputer, so groups can all work from one stable version. pyISSM lets a researcher configure, run and interact with an ISSM model entirely in Python on the same machine. The Cryosphere Community Data Pool is an open-access resource run jointly by ACCESS-NRI and the Cryospheric Sciences Community Working Group, and ccdtools is the Python package that queries and loads what sits in it.
"Our deployment of ISSM is optimized for our Australian systems, so it makes running and processing ISSM simpler and more accessible to the community," said Dr. Felicity McCormack, a co-developer of the model and chief investigator on the ARC Special Research Initiative Securing Antarctica's Environmental Future at Monash University. She said the Australian modeling community can "run the same model and data sets on the same system."
Coupling comes later: none of the four releases is a coupled ice-ocean-atmosphere configuration. The ice-sheet team has begun work on sea-level fingerprint modelling, aimed at decadal- to centennial-scale projections of global and regional sea level, and on ACCESS-AIS3, a whole-Antarctic community ice sheet model. The announcement does not include a timeline for either.
Ice sheet modelling is already a major area of research at Monash University, the University of Tasmania and the Australian Antarctic Division, and is used by ocean modellers around the country, according to Professor Nerilie Abram, the division's chief scientist. ACCESS-NRI says it will push adoption with training materials, workshops and support for the Cryospheric Sciences ACCESS Community Working Group. "We want the Australian cryospheric science community to grow, with more accessible, reliable and easier-to-use tools, so people can spend more time doing science instead of fighting with code," Tetley said.
What to watch
- Whether Antarctic results start appearing from the supported Gadi deployment instead of local ISSM builds.
- How many groups at Monash, the University of Tasmania and the Australian Antarctic Division adopt pyISSM for production runs.
- Whether the sea-level fingerprint support is delivered before the whole-Antarctic ACCESS-AIS3 configuration.