I’m not certain how the configuration information will work for rAM3 with CABLE. Is that informative to have GAL9 in the filename when it isn’t a pure GAL9 configuration?
The naming of GAL9 and RAL3P3 is what ACCESS-NRI have chosen to be the default values for rg01_rs01_m01_nameandrg01_rs02_m01_namein the vanilla release - I presume you would be changing this when CABLE is released (i.e. to GA9-CABLE?), so that wouldn’t that address that aspect?
Same thing if anyone runs a personal configuration but leaves GAL9 and RAL3P3 in the configuration file. Would we end up with a bunch of files that aren’t quite what they are saying they are?
This is currently also true with the directory naming structure. Arguably easier to fix than if files are mis-named. Perhaps the individual experiment configurations could remain as directory names, and we then just need to be very careful that nobody ever tries to consolidate all their identically-named output files into a single directory (which is the current case and hasn’t been an issue yet to my knowledge).
Ultimately I think we need to appreciate that all users will be running personal configurations, and the best way to address these risks is through clear training and documentation.
Yes, to clarify these are just the softcoded names each user gives for each region (e.g rg01), resolution (e.g. rs01) and model configuration (e.g. m01). In the case of the NRI example Lismore setup, the configurations are given names GAL9 and RAL3P3, but users can name these whatever they like through the GUI or conf file.
A typical run will have 2-3 concurrent (often nested) simulations, hence the need to differentiate their outputs with unique filenames. However it’s possible (and common) to have multiple model configurations running within the suite as we run a control and experiment(s) simultaneously. This is one of the strengths of ACCESS-rAM, its flexibility for experimentation.
To expand a bit more on what @mlipson has just said, it’s not unreasonable or unexpected that a user might have two, three, more, identically-sized and placed child nests inside a parent nest, each with a different science configuration (i.e. running multiple experiments concurrently).
In this example, for simplicity let’s say a control and two experiments in a 12km grid driven direct off ERA5-BARRA, rg01 and rs01 will be identical for all three 12km grids, and I therefore think it is of value to have the model configuration (m01) in the output file names - even when they are sitting in different output directories.