Does anyone have an idealised UM suite on gadi?

Atmosphere
1

I’m trying to get an idealised Unified Model suite running on gadi.

I think the least painful way to do this is to port a Met Office idealised (doubly periodic boundaries) UM suite to gadi, rather than attempt to convert a rAM3 / AM3 into the idealised model.

But this is a UM 13.7 Cylc 8 suite configured to run on the Met Office supercomputer. Its site file is vastly different from both a rAM3 gadi site file and a GC5 gadi site file, so it is a nontrivial undertaking.

Before I get too deep, does anyone in the community have a version of the idealised UM that is configured to run on gadi, even if it’s a very old suite?

2

If this is a different request to any extent, if anyone has access to the aquaplanet UM configuration on gadi, we would be extremely interested in obtaining that.

3

That is a completely different request but also sounds good to me :slight_smile:

Maybe post a new topic so that ‘aquaplanet’ comes up in the title for people.

4

@Michael_Barnes I couldn’t find anything in a rosie gui search for “nci gadi aquaplanet”, “nci aquaplanet” or “gadi aquaplanet”, so I’m not terribly hopeful on that front.

5

Fair enough! It was extremely wishful thinking :slight_smile:

6

@Michael_Barnes I’m currently attempting to port a Met Office idealised suite to run on gadi.

This is software engineering not science, and well beyond my knowledge, experience, expertise and comfort level :laughing:

If I succeed (see previous statement) then I might be able to try porting an aquaplanet suite.

7

BRAVE!!! Definitely would be keen to learn from you if you succeed!

8

We ran aquaplanets with um7.3 way back in the day - just need to set up ancillaries so there’s at least one land point for it to work, we put them on the poles. https://journals.ametsoc.org/view/journals/clim/33/4/jcli-d-19-0079.1.xml. That would be pre-cylc but the same idea would work with a GA run.

For idealised stuff we’ve not really done them in australia to my knowledge, just single column experiments. UKMO did have things like a hill on a flat plane. https://code.metoffice.gov.uk/doc/um/latest/papers/umdp_036.pdf

9

Thanks Scott. I thought that might be the case with the idealised stuff (I’ve run the SCM at the Bureau so I know there is a gadi port of that).

I have a UKMO idealised suite (periodic BCs in East-West), just have to get it configured on gadi…

@Paul.Gregory – see aquaplanet and land point discussion!

10

@Scott on the aquaplanet question, do you still have that GA6 suite and its aquaplanet ancil suite?

11

I do not sorry