The Australian Community Climate and Earth System Simulator (ACCESS) has been extended to include land and ocean carbon cycle components to form an Earth System Model (ESM). The current version, ACCESS-ESM1.5, has been mainly developed to enable Australia to participate in the Coupled Model Intercomparison Project Phase 6 (CMIP6) with an ESM version. Here we describe the model components and changes to the previous version, ACCESS-ESM1. We use the 500-year pre-industrial control run to highlight the stability of the physical climate and the carbon cycle. The long spin-up, negligible drift in temperature and small pre-industrial net carbon fluxes (0.02 and 0.08 PgC year−1 for land and ocean respectively) highlight the suitability of ACCESS-ESM1.5 to explore modes of variability in the climate system and coupling to the carbon cycle. The physical climate and carbon cycle for the present day have been evaluated using the CMIP6 historical simulation by comparing against observations and ACCESS-ESM1. Although there is generally little change in the climate simulation from the earlier model, many aspects of the carbon simulation are improved. An assessment of the climate response to CO2 forcing indicates that ACCESS-ESM1.5 has an equilibrium climate sensitivity of 3.87°C.

5The future behaviour of the carbon cycle is a major contributor to uncertainty in temper-6 ature projections for the 21st century. Using a simplified climate model, we show that, for 7 a given emission scenario, it is the second most important contributor to this uncertainty 8 after climate sensitivity, followed by aerosol impacts. Historical measurements of car-9 bon dioxide concentrations have been used along with global temperature observations to 10 help reduce this uncertainty. This results in an increased probability of exceeding a 2 C 11 global-mean temperature increase by 2100 while reducing the probability of surpassing 12 a 6 C threshold for non-mitigation scenarios such as the SRES A1B and A1FI scenarios 1 , 13 as compared to IPCC projections. Climate sensitivity, the response of the carbon cycle 14 and aerosol effects remain highly uncertain but historical observations of temperature 15 and carbon dioxide imply a trade-off between them so that temperature projections are 16 more certain than they would be considering each factor in isolation. As well as pointing 17 out the promise from the formal use of observational constraints in climate projection, 18 this also highlights the need for an holistic view of uncertainty. 19Introduction 20The last decade has seen a great deal of research into quantifying the uncertainty of climate 21 projections as a guide to sensible adaptation and mitigation activity. The two most common 22 techniques have been multi-model ensembles (such as the CMIP3, Third Coupled Model In-23 tercomparison Project 2 and C4MIP, Coupled Carbon Cycle Climate Model Intercomparison 24 Project 3 ) and perturbed physics experiments 4 . These experiments have allowed us to isolate 25 those factors responsible for the spread of model ensembles. They have identified factors con-26 tributing to the overall climate sensitivity and uncertainty in the radiative response to aerosol 27 changes, as well as the response of the carbon cycle, including some assessment of the feed-28 backs between climate change and carbon fluxes. 29 More recently, observations have been used to provide direct constraints on model behaviour 5, 6 . 30 Computational complexity has largely precluded the use of full Earth system models (ESMs) 31 and restricted the facets of model behaviour that can be constrained. Full ESMs generally ex-32 plore fewer model parameters or use statistical emulation to expand their sample size. Most 33 critically, the carbon cycle behaviour remains largely untreated (with a rare exception 7 ), despite 34 the sensitivity of global-mean temperature change projections to carbon cycle processes 8 . In-35 stead, typically, simple models have been calibrated against the coupled carbon cycle/climate 36 models of Friedlingstein et al. 3 for which only limited formal calibration is possible and which 37 do not address parametric uncertainty. 38 Here we apply historical observations to address the scientific uncertainty of the climate 39 system. Our general method has been to identify those mo...

A new version of the Australian Community Climate and Earth System Simulator coupled model, ACCESS-CM2, has been developed for a wide range of climate modelling research and applications. In particular, ACCESS-CM2 is one of Australia's contributions to the World Climate Research Programme's Coupled Model Intercomparison Project Phase 6 (CMIP6). Compared with the ACCESS1.3 model used for our CMIP5 submission, all model components have been upgraded as well as the coupling framework (OASIS3-MCT) and experiment control system (Rose/Cylc). The component models are: UM10.6 GA7.1 for the atmosphere, CABLE2.5 for the land surface, MOM5 for the ocean, and CICE5.1.2 for the sea ice. This paper describes the model configuration of ACCESS-CM2, documents the experimental set up, and assesses the model performance for the preindustrial spin-up simulation in comparison against (reconstructed) observations and ACCESS1.3 results. While the performance of the two generations of the ACCESS coupled model is largely comparable, ACCESS-CM2 shows better global hydrological balance, more realistic ocean water properties (in terms of spatial distribution) and meridional overturning circulation in the Southern Ocean but a poorer simulation of the Antarctic sea ice and a larger energy imbalance at the top of atmosphere. This energy imbalance reflects a noticeable warming trend of the global ocean over the spin-up period.