Article
BAMOS February 2026
15
Given that every model approximates the real world, the robustness of future climate projections is increased by considering a multi-model ensemble of simulations. Over many decades, the World Climate Research Program has coordinated CMIP to provide that multi-model ensemble. Australia has participated in all previous CMIP phases, with CSIRO leading model submissions.
Simulations with ACCESS-ESM1.6 are currently underway to contribute to the CMIP7 Assessment Fast-Track— a set of prioritised experiments which will deliver early results to be used in the upcoming IPCC Seventh Assessment Report. IPCC Assessment reports are vital for climate policy development under the United Nations Framework Convention on Climate Change.
In addition, ACCESS-ESM1.6 will contribute to a number of CMIP7 Community MIPs, which address a wide range of scientific questions through community-led domain-specific experiments without a strict timeline. For example, the new model version will be used to explore carbon cycle feedbacks and carbon dioxide removal strategies.
Advancing Australian climate research
ACCESS is extensively used by the Australian community in many areas of research. Future climate projections based on ACCESS and other products, such as simulated carbon stocks, are also widely used by Australian agencies and the government.
ACCESS-ESM1.5, Australia’ s contribution to CMIP6, already provided large ensemble simulations of up to 40 members, enabling us to quantify uncertainties and provide probabilistic information to policy and decision makers for risk assessments, model evaluation, and exploring low-probability high-impact events.
The model also enabled researchers to design and perform unique experiments, including long-term climate stabilisation simulations to explore the impact of net-zero emissions on regional climate and extremes and to evaluate the role of the Australian terrestrial biosphere under different future scenarios.
ACCESS-ESM1.6 will continue to provide this capability to the Australian community, with significant upgrades and new features, as well as an improved performance of up to 25 simulation years per day on the National Computational Infrastructure( NCI) supercomputer, where ACCESS is set up and run.
Looking ahead, maintaining and further developing ACCESS- ESM will be essential to ensure it remains world-class, fit for purpose, and policy-relevant. Achieving this will require multidisciplinary skills and expertise across organisations.
Figure 1: Comparison of forest type area coverage( four left-most panels) and the total gross primary productivity( right-most panels) between ACCESS-ESM1.5( top row) and ACCESS-ESM1.6( bottom row). ACCESS-ESM1.6 includes two new Australian-specific forest types – mesic broadleaf forest and xeric broadleaf forest – consistent with independent, nationally-consistent datasets, resulting in changes to simulated vegetation carbon uptake( averaged over the period 2000-2015). Credit: Alex Norton