As part of my work generating atmosphere ancillaries for our ACCESS-AM3 model, I re-wrote the ozone script which was in the u-dj813 suite to be more (in my eyes) logical and understandable. The task and script are here.
The repository is still private- I haven’t done a detailed check of the history to ensure there’s nothing about UK machines remaining. Also, while all current paths in the workflow point to locations on our local HPC Gadi, I assume they would reflect directory structures on UKMO machines. We are yet to decide how this ancillary workflow and the CMIP7 workflows should interact, as some of their goals are overlapping, but not all.
Plots of ozone with the old script vs new:
Looks like I’m off by some small scaling factor, which I’m trying to track down.