COSIMA
Identify Core Datasets needed to run the recipes
While working on the CI/CD and on the conversion of the recipes to use the intake catalog, I noticed that a lot of data is stored in hh5 and ik11.
I propose that the MED team put together a list of the datasets needed to run the recipes. We can then discuss a plan to consolidate and support the data.
Good idea to think about this @rbeucher. In many cases it is somewhat arbitrary what data is used in the recipes, and in some we've tried to use lower res model output so that the recipes run faster.
If ACCESS-NRI has the capacity to store and maintain some of the COSIMA data on ik11/ol01/cj50 that would be great. Though perhaps it is more useful to think about which data is used most by the community, rather than what data is currently used in the COSIMA recipes. It would be easy to switch simulations in the recipes where needed.
@max-anu, @paigem should be able to help with this.
@adele-morrison, I can see there are some re-gridded WOA datasets. Is it something that would be worth adding to the ACCESS-NRI replicas collection for model evaluation? We could create a COSIMA sub-collection and start adding observation datasets.
See here for what we currently have for ESMValTool and ILAMB: https://geonetwork.nci.org.au/geonetwork/srv/eng/catalog.search#/metadata/f7199_2480_5432_9703
@rbeucher Happy to help with this! I can go through and make a list of the datasets currently used in the COSIMA recipes.
@adele-morrison I may reach out to you once I have the list to provide input on which datasets could be updated to be more useful to the COSIMA community.
@rbeucher WOA is definitely a useful one to have in the COSIMA sub-collection of obs datasets for model evaluation. In the past we've found it useful to have this pre-interpolated onto the various ACCESS-OM2 resolutions, so people don't have to redo the interpolation themselves all the time.