‘Run status’ is given as NxM yr where N is the number of ensemble members that have been completed and M is the number of years required for the submission (some experiments may have run a few extra years).
Raw model output is available here: /g/data/p73/archive/CMIP7/ACCESS-ESM1-6/production in the directory listed under ‘p73 location’. Access to p73 can be requested through MyNCI. The piControl run is split across two directories. The first 100 years in the esm-piControl directory is considered spin-up.
The historical experiments run from 1850 to 2021. Scenarios will run from 2022 to 2100 with forcings from integrated assessment models and to 2150 or longer with more idealised forcings.
Concentration-driven idealised simulations: global mean surface air temperature
The temperature anomaly is calculated as the difference from the mean of the first 100 years of the piControl simulation.
The control simulation (run under constant 1850 conditions) (black) shows stable global mean temperature as expected. Temperature increases rapidly (red) when atmospheric CO2 is instantaneously quadrupled. The temperature increase is approximately linear when CO2 is increased by 1% per year (blue). In the case where only the biogeochemistry is forced with the increasing CO2 (green), the temperature is close to that of the control run. In the case where only the radiation scheme is subject to the increasing CO2, the temperature increase is close to that of the standard 1pctCO2 case.
Concentration-driven idealised simulations: carbon fluxes
The figure shows the globally integrated carbon source (positive) or sink (negative) to the atmosphere from the land (thin lines) and ocean (thick lines). A 10-year running-mean is shown as the land fluxes have large interannual variability responding to interannual climate variability.
The control simulation (black) shows carbon fluxes that are close to zero as required. Land carbon fluxes are more variable than ocean carbon fluxes. When atmospheric CO2 is instantaneously quadrupled, both land and ocean take up carbon (red) with the ocean sink being stronger and lasting longer than the land sink. When CO2 is increased by 1% per year, both land and ocean initially respond by taking up carbon (blue) but the land sink reduces over time and becomes a source as temperature increases. The behaviour is similar when only the biogeochemistry is forced with the increasing CO2 (green) with both ocean and land showing slightly larger CO2 sinks. When only the radiation scheme is forced with increasing CO2 (magenta), the ocean carbon flux remains close to zero while the land shows a carbon source, driven by the temperature increase in this case.
Emissions-driven idealised simulations
In emissions-driven simulations the experiment is forced with anthropogenic emissions rather than atmospheric CO2 concentration. Simulated land and ocean carbon fluxes change the atmospheric CO2 concentration.
The figures show the emissions-driven piControl (black) and esm-flat10 (red). The esm-flat10 experiment has globally and temporally uniform emissions totalling 10 PgC/yr. The CO2 concentration in the esm-piControl is very stable (as desired) while it grows approximately linearly in the esm-flat10 case but more slowly than the concentration-driven 1pctCO2 (blue). The temperature increase is also slower than 1pctCO2. A larger maximum temperature anomaly is reached in esm-flat10 despite a lower maximum CO2 concentration. This indicates a feedback between the carbon cycle and climate.
The land and ocean carbon fluxes respond to the increasing atmospheric CO2 and changing climate. The ocean flux is mostly driven by the change in CO2 and responds with much larger uptake in the 1pctCO2 case than in the esm-flat10 case where there is more time for ocean CO2 to start to equilibrate with the atmosphere. The land carbon fluxes are more similar suggesting a larger influence from the warming climate.
Emissions-driven historical run
The emissions-driven historical experiment runs from 1850 to 2021 and is forced by fossil fuel carbon emissions. Atmospheric CO2 depends on these emissions and modelled carbon uptake by the land biosphere and ocean. The impact of carbon in the atmosphere due to land-use change is also simulated based on prescribed changes to the mix of vegetation types in each grid-cell.
Global mean simulated atmospheric CO2 for the lowest model level (left panel, red) compares well with the whole-of-atmosphere atmospheric CO2 (black) that would be prescribed in a concentration-driven historical simulation. Note that atmospheric CO2 is relatively well-mixed through the troposphere on annual timescales. Ensuring a good match with atmospheric CO2 at the end of the simulation has been prioritised over the earlier part of the historical period, to ensure that future scenarios are starting from current atmospheric CO2 levels.
Global mean surface air temperature (right panel, red) compares well with observed temperature (black). Interannual variations (other than those externally forced e.g. volcanoes) are not expected to match between the model simulation and observations. An ensemble of historical experiments will be run to better characterize internal variability and forced changes.
