Code management group minutes 4Dec2012
UKCA Code Management Meeting, 4 December 2012 (tele-conference).
Met Office: Nick Savage (NS), Colin Johnson (CJ), Mohit Dalvi (MD)
Cambridge: Alex Archibald (AA)
Leeds: Graham Mann (GM)
Oxford: Zak Kipling (ZK)
Not present but on CMG; Sandip Dhomse (Leeds, SD), Luke Abraham (Cambridge, LA), Paul Telford (Cambridge, PT).
Summary of action points arising in the meeting
- NS to report back on how we make current tickets available to those outside the Met Office.
- Reaction rates to be updated for CheST configuration only.
- GM to discuss nitrate testing with Steve.
- MD to finish tying evaluation suite together.
- MD to continue testing UKCA/ENDGAME tracer tranport.
- GM to look at tuning scavenging @ 8.2.
- CJ to get 2-mode dust scheme working.
- NS to set up Doodle poll for Jan meeting.
Carried forward from previous meetings
- GM to collate UKCA Science plans.
- CJ to circulate provisional list of UKCA configurations at/post UM8.4.
- ZK to request Philip to circulate information/updates on JASMIN.
- MD to update PUMA 8.2 TropIsop+MODE job with latest changes.
For beyond next meeting
- New reaction rates to be tested in CheS and CheT.
- CJ to merge in-plume convective scavenging at 8.2, with help from ZK.
- CJ to lodge offline-oxidant code.
Status of previous actions
- NS to check admin on UKCA email lists. JISCmail – look like only Luke can do this.
- NS to circulate list of Somewhen tickets on UKCA. Not copied across yet. This is a tedious process, and hard to keep up to date. We clearly need better collaboration tools. NS to look into this and report back.
- GM to collate received UKCA Science plans into a document. Not collated yet. GM to do so before Christmas to inform JCWRP report (early Jan) and board meeting (Feb?)
- CJ to circulate provisional list of UKCA configurations at/post UM8.4. Still to do.
- ZK to request Philip to circulate information/updates on JASMIN. Still to do.
- NS to circulate link for external release of Iris postprocessing package, with info in UKCA/Iris experience. Unfortunately the AQUM UKCA/Iris project has stalled due to funding issues. It is hoped to pick this up again, but at present it is poorly documented. Iris itself is publicly available, with a Google group for discussion.
- CJ/MD to port a UM8.2 TropIsop+MODE job to PUMA including latest improvements/fixes. UKCA @ 8.2 (TropIsop+MODE) is on PUMA as of mid-Nov. Update to come this week, based on
- NS: Little progress on tracer rationalisation, but a better understanding of the current situation. We'll need a new tracer for sodium, and to clarify the sea salt tracer, before nitrate goes in.
- NS: Fiona has tested new reaction rates in CheST @ 8.2, but what do we do about CheS and CheT (where the new rates have only been tested at 7.3)? Not enough resources to test these @ 8.2 at present. It would be good if rates were switchable, but not possible yet. Decision: Only update CheST for HadGEM3 ESM1, but prioritise testing CheS/CheT in the new year.
- Steve is teting the nitrate code. GM to discuss this with him. However, this is no longer slated for 8.4, but will be in 9.0 (which was previously to be called 8.6).
- MD: Evaluation suite code is on
postproc03, but still being tied together
- MD: Currently checking tracer transport in UKCA with ENDGAME. Refer to experience with CLASSIC tracers.
- CJ: Nucleation-scavenging timescale in 8.2 longer for SU/SS, but shorter for OC, compared to 7.3. Large-scale now looks good, but convective problematic. Will try to merge ZK's code in Jan. GM to look at tuning the existing scheme.
- CJ: Currently trying to get a 2-mode dust-only scheme working – problems in
UKCA_EMISSION_CTLdue to dependencies on chemistry.
- CJ: offline oxidant configuration giving reasonable results, may lodge in January if time. Still not that fast (with N-R solver), so need to try different solvers (B-E, KPP).
- ROSE progressing well.
- 8.4 frozen end Jan.
- 8.5 closed to bring in ROSE support (UMUI still supported too at this point).
- 9.0 first ROSE-only release, will also include science changes.
- ROSE will be used operationally by September.
- Richard Barnes has the ticket for UKCA ROSE support, but will need help from the rest of us to understand the current control logic.
Date of Next meeting
Mid-January, NS to setup doodle poll (http://www.doodle.com/8aacaepan75mrugr).