Hi team. I plan to run ACCESS-AM3 AMIP-style N96 simulations from 1982-2024. I realised that some ancillaries only date to around 2014, such as ozone (2015), seaice (2012), sst (2012), volcanic aerosols for easyaerosol (2014), UKCA ancilarries (2015), see install_ancil->file in the alpha release of ACCESS-AM3.
Would it be possible to get support on how to get and specify ancilarries for more recent years, ideally up to 2024/25 for better comparisons with latest observations.
I just found newer sst files up to 2019 here: /g/data/vk83/prerelease/configurations/inputs/access-cm3/ancil/2025.07.25/n96e/mom025_20250315/sst/reynolds/1981_2019_360day/qrclim.sst.
Do we just need to replace the file name?
Again, is it easy to extend beyond 2019?
Short answer is not yet, but we do plan to update the ancillary sources to more up-to-date versions for the full release.
I suspect the answer for each of the ancillaries are different, but the long term plan is to support all of the time-dependent ancillaries up to as recent a date as possible. The difficulty will vary between ancillaries- some are still being generated in the same form, which would be easy to substitute in, and some others not.
If you want to inspect how each of the ancillaries are created, take a look at Martin’s ancillary suite (u-dj813). It was the suite used to create the current n96 ancillaries, and the n512 ones being used for the high-res runs. We are working on improving the land part of the ancillaries for the Beta release, but the atmosphere/ocean ancillaries are likely to remain the same (although hopefully with updated input files).
Hi Qinggang!
In my ACCESS-AM2 simulations for 2015-2024 I use ssp245 emission estimates for ozone, BC, SO2 and OC. I’ve also created sea ice and SST ancillary files for 2015-2024 based on modified MetOffice processing code using the HADISST dataset for a n96 grid (this took me quite a while…but I’ve written up some documentation on how to do this so shouldn’t be too hard if you need to expand to 2025!) - although note the 2024 data for sea ice/SSTs contains uncorrected data (I can talk to you more about this). The other ancillary files you mention are also likely in my rose config files somewhere…
@alanah.chapman Could you please share me the documentation about how to create my own SST/Sea Ice. Very likely I will go for ESA CCI SST/Sea Ice, but I think the procudure should be similar.
Note that for this method you need monthly data, and the first and last years of the dataset need to be full years (so if you want to extend to 2025 then you need to either wait until the dataset has December values OR use December 2024 values but don’t run the model for the year 2025). The 2025 SST-sea ice ancillary data will be uncorrected using the MetOffice code (because a full year of data is needed to correct the previous years data) - this means that the 2025 SST-sea ice data won’t match the monthly means as accurately as the rest of the ancil dataset (I have links in the repo to the MetOffice where they explain in more detail).
I had to edit the original fortran code from the MetOffice, but I don’t think I can publicly share the edits I have made because I am not the original author…and the MetOffice code needs a MOSRS account to access…but I think it should be fine to tell you the changes I made → these are in a private repo which is linked to the above public repo.