Set up AMIP config

[blimlim]

Coecms has an amip configuration of ESM1.5 available in a separate repository to the other ESM1.5 configurations: https://github.com/coecms/access-esm-amip

It would be good to bring this in line with the already released configurations and make it available. This would involve:

• Cleaning up the configuration files, removing any that aren't needed
• Adding netCDF conversion
• Adding diagnostic presets
• Moving ancillary files: SSTs and sea ice concentrations onto vk83
• Updating the config.yaml to use the newer payu features.
• Adding the CI/CD infrastructure
• Add land use change script

The existing configuration does not contain any land use updates. We should check whether these were included in the simulations done by CSIRO.

[blimlim]

I've created a working branch here which I'll turn into a dev-amip+concentrations branch once the basic changes are in.

I've swapped over to the NRI executables on vk83. Interestingly the coecms version used a different executable path /g/data/access/payu/access-esm/bin/coe/amip/um7.3x to the historical and pre-industrial configurations /g/data/access/payu/access-esm/bin/coe/um7.3x. These are identical though:

$ cmp --silent  /g/data/access/payu/access-esm/bin/coe/um7.3x /g/data/access/payu/access-esm/bin/coe/amip/um7.3x && echo "same" || echo "different"
same

and so I think it's ok for us to use our standard executable without any modifications.

@MartinDix Noted that the CSIRO amip runs included the same land use changes as the historical simulations:

(comparing /g/data/fs38/publications/CMIP6//CMIP/CSIRO/ACCESS-ESM1-5/amip/r1i1p1f1/Lmon/landCoverFrac/gn/latest/landCoverFrac_Lmon_ACCESS-ESM1-5_amip_r1i1p1f1_gn_197901-201412.nc and /g/data/fs38/publications/CMIP6/CMIP/CSIRO/ACCESS-ESM1-5/historical/r1i1p1f1/Lmon/landCoverFrac/gn/files/d20191115/landCoverFrac_Lmon_ACCESS-ESM1-5_historical_r1i1p1f1_gn_185001-201412.nc)

The land use change script is absent in the coe configuration, and so we will need to add this in.

I've also restructured the files, moving all the UM config files into an atmosphere directory. This means that output files will also be written to an atmosphere subdirectory. I think this makes the config a bit cleaner, though the main reason I did this is admittedly due to an oversight I had when writing the ESM1.5 conversion driver for um2nc which assumes this directory structure.