access_ACCESS_AMIP_experimental - ACCESS-NRI/accessdev-Trac-archive GitHub Wiki
Experiments based on the standard ACCESS AMIP configurations
This page describes some possible modifications to the standard ACCESS AMIP configurations. They're intended as background for the CWSLAB testcases and to illustrate the range of things that might be done with the model.
The standard CMIP5 AMIP configuration (both ACCESS1.0 and 1.3) used
- Monthly SST and ice data for 1979 to 2008
- Annually varying greenhouse gas conentrations and emissions of aerosol sources
- Monthly and annually varying ozone (so including effect of ozone hole)
Possible sensitivity experiments could include any of the following
- Climatological or pre ozone hole ozone
- Climatological SSTs and ice to look at effect of interannual variation
- Constant GHG
- Starting from a perturbed initial state to create an independent ensemble member
- Starting from an modified initial state (e.g. no soil moisture) to find how long the model takes to equilibrate
Modifications to the model physics configuration could include
- Turn off gravity wave drag
- Changing CAPE closure scheme or timescale in the convection
Simple changes to the configuration could include
- Requesting additional model diagnostics
More details
Model diagnostics
Model diagnostics are controlled by the UM Spatial and Temporal Averaging and Storage Handling (STASH) system which has its own section of the UMUI. For more details see the UM user guide https://accessdev.nci.org.au/umdocs/7.3/UM_docs/papers/html/000/node34.html
In full CMIP5 configuration the standard AMIP runs save a large set of monthly mean diagnostics to the 'pa' files, daily surface, TOA and limited pressure level diagnostics to the 'pe' files, 6 hourly data atmospheric data to the 'pi' files and a small set of 3 hourly fields to the 'pj' files.
Some important variables like surface air temperature are saved to multiple output streams.
Collections of variables for the various CMIP5 tables are set up as packages within the STASH section of the UMUI and can be turned on or off in a group.
Ozone
The model normally uses an ancillary file CMIP5 /projects/access/data/ancil/SPARCO3_1850-2009_L38.anc. This ozone data is a combination of observations and modelling. It represents the Antarctic ozone hole, though not the interannual fluctuations. This graph shows the October ozone at the South Pole at the model level of maximum decrease (level 30).
A possible experiment is to run the model with fixed 1960 ozone. For this the ancillary file should be replaced by /projects/access/data/ancil/SPARCO3_1960_L38.anc. The screenshot of the UMUI shows where to find this option
After making this change you can use the difference feature of the UMUI to compare the new job and the standard job. This should show the difference in the ozone file names.
The ozone concentration is nor normally saved as a model diagnostic but it is saved in the model restart files so it's possible to plot the difference in ozone fields from these.
