HadGEM3.1 N96-ORCA1 @vn10.7

[ajd]

Hello,

I have heard that the CMIP6 HadGEM3.1 version has been ported to Monsoon and Archer. Is it currently possible to use this suite and if so what is the suite-ID?

Also, is there a 'map' for how model output can be related to CMIP variables? For example, which STASH codes would I need to get vertically integrated liquid water path? So far I've been working with STASH 1242 and 1223.

Many thanks!
Andrea

[ros]

Hi Andrea,

Yes the suites for Monsoon (u-ar930) & Archer(u-as037) are detailed here along with a link to the Met Office suite webpage: http://cms.ncas.ac.uk/wiki/UM/Configurations/GA7.0-GC3.0

The Archer suite does not yet have post-processing to RDF added in yet. I'm hoping our changes will make it onto the trunk very soon (currently in review) and I would like to wait until that happens if at all possible.

There is a link from the Met Office webpage for suite u-ar766 that details CMIP6 diagnostic mappings - does this answer your question? ​https://code.metoffice.gov.uk/trac/ukcmip6/wiki/ExperimentGuidance

Regards,
Ros.

[ajd]

Hi Ros,

Great, many thanks for this! We will most likely run on monsoon, is that suite ready to be used?

Thanks for the link, this is very useful!

Cheers,
Andrea

[ajd]

Hi again Ros,

Further to this, I'll wait to use the suite properly until the post-processing has been added, but I'm planning to try running a test today - is this ok?

Also, what is the situation with disk space on the RDF and is there a quota per user? Do we have options for long-term storage on tape at BADC?

Many thanks,
Andrea

[grenville]

Andrea

The PI of the project you are working on might be better placed to answer the BADC question - what is the name of your project? We strongly encourage use of elastic tape - it does provide long-ish term storage, but it is not an archive (it is not curated).

RDF space is available - could you provide an estimate of requirements?

Grenville

[ajd]

Hi Grenville

Thanks for your reply. I'll discuss further with Ed Hawkins, the project is SMURPHS.

We'll be running ~28 historical runs with HadGEM3.1, so while we don't know the exact amount of storage needed yet, it will likely be around 300TB (that was our estimation for the related REAL projections project which will continue this ensemble into the future).

Can we add Ed to this ticket? I'm not able to do this unfortunately.

Many thanks,
Andrea

[grenville]

Andrea

There is already a SMURPHS group workspace space on JASMIN - the group in Leeds manage that (and the ARCHER SMURPHS account). Please contact Chris Smith (c.j.smith1@…) for for information.

Grenville

[ajd]

Thanks Grenville,

I didn't realise the group workspace was for storing the model output - but that is for temporary storage only, is that correct? Also, since we will be running on monsoon, is that where I am meant to move the model output to?

I also have couple of non-archiving related questions:

I tried running a month from u-ar930 as a test, and it worked! Thanks. However, I get a very different view of the STASH list than for similar suites on puma, and the error: Could not find metadata for /data/users/moci/CMIP6_rose-meta/um-atmos/HEAD - is this supposed to happen?

Also, I noticed the suites are PI control runs - will there be a similar suite for historical runs or is it best to modify this suite myself to be consistent with the MO historical suites?

Many thanks and apologies for the many questions!
Andrea

[grenville]

Hi Andrea

I think we discussed these things at the UM training. I'll close this ticket now, but reopen if that's premature.

Grenville