Minutes of the code management group meeting 2015-09-18
Present
Nick Savage (NS), Fiona O'Connor (FOC), Alan Hewitt (AH), Nicolas Bellouin (NB), Mark Richardson (MR), Jane Mulcahy (JM), Zak Kipling (ZK), Alex Archibald (AA)
Apologies: Luke Abraham (LA), Mohit Dalvi (MD), Colin Johnson (CJ), Graham Mann (GM)
Actions from this meeting
One off actions
- (ongoing) JM/GM to talk offline regarding improvements.
- (ongoing) JM to open ticket. JM to check which ticket this referred to.
- (ongoing) CJ to email LA re problems with SO2 diagnostics for emissions. Issues related to using ASAD diagnostics package.
- (ongoing) CJ to put branch on PUMA email CMG location of branch to check details.
- (ongoing) GM check 8.4 versus 7.3 GLOMAP-mode code and report any bug fixes the bugfix page.
- (ongoing) AA GLOMAP-mode coupling to FJX. AA to see if Paul Griffiths can be asked to do this.
- (ongoing) Incorporate Bug #35 into dry deposition work - Federico Centoni (FC) wants to do some sensitivities - hold off until FCs assessment - still waiting.
- (ongoing) ALL to migrate tickets over to the shared repository (MOSRS). Some people have done this, but not all ticket migrated yet.
- (ongoing) NS to continue with lodging ticket #368 quick fix. A proposed fix to the value of RAFEPS allowing just one line to have a scaling factor on a namelist has been developed but NS is concerned about implementing it it due to his lack of understanding the code. Implementation will be pending the next action.
- (ongoing) LA to contact Oliver Wild regarding NR solver. In particular to ask advice on the use of the increased value of rafeps to prevent model failures. How safe is this, and is there a better solution?
- (in progress) Bug #38: GM to take on lodging this on the trunk.
- (ongoing) JM to contact Alistair Sellar and check on criticality of bug #38 (volcanic emissions) for UKESM1.
- (ongoing) LA to check doc for NVOC factor and open ticket to remove it from code.
- (ongoing) FOC to raise with Ben Shipway (as well as MH from previous LFRic meeting) regarding when aerosols will be needed for LFRic.
- (ongoing) FOC to raise with David Walters about aerosol climatologies for LFRic. '
- FOC to follow up about Rose stem on MONSooN.
- JM to put the branch to fix nucleation diagnostics in the microphysics on the bug fixes wiki together with a list of changes need to aerosol tunings.
- JM to prepare some slides on the impact of the bug in the microphyics on aerosol and upload to this wiki.
- NB and James to look at the impact of the bug in the microphyics and the tunings needed to go with it on performance in a GA4 job.
- NB to add branch for GA4 job to UKCA wiki bug fixes table.
- NS to add owner and CLOSED/OPEN columns in bugs table.
- Ben Johnson (BJ) to review OC biomass burning emissions units (#41 in bug table) and if this is now correct, close.
- NS to investigate issues with aerosol surface area (#52) and find more suitable owner for ticket.
- MR to provide information on his fixes to RADAER files on the bugs table (#53).
- JM to investigate current status of netCDF files in coupled jobs (#66)
- NS to open ticket to capture issue with time availability codes for UKCA chemistry timesteps. Talk to Jane about her requirements and issues. See if CRUM team will take on the work.
- NS if issue with output of NTPs on pressure levels not resolved before lodging the code, open a ticket to investigate further after 10.3.
- FOC to open a doodle poll for December and 2nd Week January.
Standing actions
- (ongoing - standing item) CMG to monitor list of bugs and circulate between group for owners.
- (ongoing - standing item) At CMG future meetings there should be included an appendix listing the open bugs. If a ticket is included it is not an open bug (as the owner of ticket is person doing lodging).
- (ongoing - standing item) JM to continue to link/send powerpoint results (re: GA7.0), as new configurations released and link/put files on UKCA wiki.
- (ongoing - standing item) ALL with access to keep monitoring ukca_bugs email list.
- (ongoing - standing item) ALL to think about if code changes will change KGO and let systems team know in good time. This should be a standard part of our working practice.
Teleconference Numbers
- UK Freefone: 0800 9171956
- Participant passcode: 89264837 then #
Agenda
- Actions from the last meeting
- GA7.0 update
- UKCA Rose suites
- Rose/Rose stem on MONSooN
- Bugfixes
- Targets for vn10.3
- LFRic
- A.O.B./D.N.M.
- Date of next meeting
Minutes
Actions from 2015-05-06
- JM to circulate GA7.0 email. Done
- JM/GM to talk offline regarding improvements. Ongoing
- JM to open ticket. Ongoing - JM to check.
- (ongoing) CJ to email LA re problems with SO2 diagnostics for emissions. Issues related to using ASAD diagnostics package. Ongoing
- (ongoing) CJ to put branch on PUMA email CMG location of branch to check details. Ongoing
- (in progress) GM check 8.4 versus 7.3 GLOMAP-mode code and report any bug fixes the bugfix page. Ongoing
- (in progress) Alistair Sellar to add in status and UKESM1 ticket numbers to each item (re: UKESM1 strategy document). Done
- (ongoing) FOC/NB/GM - discuss (9) offline (re: GLOMAP-mode coupling to FJX) - Apostolos Voulgarakis could also be included in this. Could this also link to the heterogenous chemistry task? . Ongoing Agreed that Paul Griffiths to be asked to do this.
- (in progress) ALL - feed in to Alistair Sellar list of priorities. Items need an owner if people need it raised. Done
- FOC to raise Rose stem with UM Systems Team, then LA may raise at UMPB or FOC with JWCR. Done. UM systems team supportive. FOC to follow up.
- LA to contact MONSooN team about Rose stem on MONSooN. Done
- (not done) Incorporate Bug #35 into dry deposition work - Federico Centoni (FC) wants to do some sensitivities - hold off until FCs assessment - still waiting. Ongoing
- FOC to follow-up with FC. Done. FC still evaluating impacts when the bug fix is applied to all species and not just ozone.
- (in progress) Bug #36 (ozone): FOC/Steven Hardiman to take on - see MOSRS ticket #102 (UKCA Bugfix for ozone column diagnostic calculation) Done
- NS to discuss wet_levels with Glenn Greed. Done. Note that since New Dynamic (UM6.0 onwards) models have been unable to run with wet_levels not equal to model levels.
- NS to possibly add-in run-time check for wet_levels for vn10.2, or remove from namelists and set to model_levels(?) Done. Wet levels has been removed from the model as of 10.3 by UM systems team.
- (ongoing - standing item) CMG to monitor list of bugs and circulate between group for owners. Ongoing
- (ongoing - standing item) At CMG future meetings there should be included an appendix listing the open bugs. If a ticket is included it is not an open bug (as the owner of ticket is person doing lodging). Ongoing
- (ongoing) ALL to migrate tickets over to the shared repository (MOSRS). Some people have done this, but not all ticket migrated yet. Ongoing
- (in progess) JM to link/send powerpoint results (re: GA7.0), updated to latest configuration and link/put files on UKCA wiki. Done. JM to continue to update this as new candidates are evaluated.
- (in progress) MD to update suites to vn10.1. Done. 10.2 suites are also available. (see later)
- NS to continue with lodging ticket #368 quick fix. A proposed fix to the value of RAFEPS allowing just one line to have a scaling factor on a namelist has been developed but NS is concerned about implementing it it due to his lack of understanding the code. Ongoing. Implementation will be pending the next action.
- LA to contact Oliver Wild regarding NR solver. Ongoing
- (in progress) Bug #38: GM to take on lodging this on the trunk. Ongoing
- JM to contact Alistair Sellar and check on criticality of bug #38 for UKESM1.
- FOC to check documentation for emissions units. Done
- (in progress) LA to check about NMVOC in new emissions routine. Done
- LA to check doc for NVOC factor and open ticket to remove it from code. Ongoing
- (ongoing) ALL with access to keep monitoring ukca_bugs email list. . Ongoing
- (ongoing) ALL to think about if code changes will change KGO and let systems team know in good time.. Ongoing. This should be a standard part of our working practice.
- LA to wait until after Exec on 2015-05-08 to contact Jeremy Walton regarding vn7.3 on MONSooN Cray. Done. UM7.3 is available on Monsoon-cray.
- NS to contact Steve Mullerworth for information on Limited Area LFRic and UKCA should start to work with it. Done. Nothing happening in the short term on LAMs and so too soon for us to engage over this.
- FOC to raise with Ben Shipway (as well as MH from previous LFRic meeting) regarding when aerosols will be needed. Ongoing
- FOC to raise with David Walters about aerosol climatologies. Ongoing
- LA to open a doodle poll for 18th June to end of June.. Done
GA7.0 update
JM reported on GA7 progress. The main work done since the last meeting has been in response to the discovery of a bug in the nucleation scavenging diagnostics which are inputs to UKCA. These were found to be approximately x10 too low with major impacts on aerosol once these were correct. A package of new tunings has been carried out in response to this, mostly adding a paramterisation to account the fraction of a gridbox which is raining and changes to impaction coefficients from snow. This has led to an overall improvement in performance but there is a a negative impact on TOA radiation. Actions resultiing:
- JM to put the branch to fix nucleation diagnostics in the microphysics on the bug fixes wiki together with a list of changes need to aerosol tunings.
- JM to prepare some slides on the impact of the bug in the microphyics on aerosol and upload to this wiki.
- NB and James to look at the impact of the bug in the microphyics and the tunings needed to go with it on performance in a GA4 job.
- NB to add branch for GA4 job to UKCA wiki bug fixes table.
UKCA Rose suites
Mohit has prepared two rose suites at UM10.2:
- UM10.2+UKCA StratTrop + GLOMAP + Rad+Activate + NetCDF emiss on MONSooN Cray --for branch testing/ development only = u-aa411 (by Jeremy Walton). https://code.metoffice.gov.uk/trac/roses-u/browser/a/a/4/1/1
- UM10.2 + GA6(ver136.17) + StratTrop+ GLOMAP + Rad + Activate + NetCDF emiss on Met Office Cray = mi-ah744. This is being assessed with a decadal run before porting to MONSooN.
Bugfixes
# | Description | Ticket | Affected Versions/ Config |
Branch with fix | Remarks | |
---|---|---|---|---|---|---|
MetO | NCAS | |||||
31 | FP exceptions in asad_sparse_vars or asad_spimjmp |
#5616, #368 (MOSRS) | #555 | 7.3: CheT+Aero |
MetO: fcm:um_br/dev/fris/vn8.4_increase_rafeps Puma: fcm:um_br/dev/nhsavage/vn8.4_increase_rafeps |
Paused pending LA discussion with Oliver Wild |
35 | Error in labelling: Surface resistance <--> stomatal resistance ? | #6417 | #653 | 7.3: | None | Awaiting input from Federico Centoni |
38 | Missing logical for volcanic so2 emissions. | #6418 | #656 | MetO: fcm:um_br/dev/hadna/vn8.6_deactivate_volcanoes Puma: None |
GM. To be Lodged | |
41 | OC biomass burning emissions - incorrect conversion. Assumes emissions as kg POM, but (new) ancils are kg C |
#6429 | vn8.5: | MetO: fcm:um_br/dev/fpbj/vn8.6_ukca_mode_BBOC-POM_bugfix Puma: None |
To be lodged Need to determine which ancils are POM vs C. BJ to review. | |
46 | Inconsistent use of wet_levels and model_levels in UKCA_AERO_CTL 1-D arrays |
None | All versions | None | Impact if wet_levels/= model_levels, but this option has never been supported for New Dynamics or ENDGAME and wet_levels has been retired by #425 for 10.3. | |
49 | UKCA emissions: incorrect number, un-initialised slots in all_emissions() |
#381 (MOSRS) | All UKCA | Example fix here (MOSRS) | To be lodged at vn10.2 as part of #805 (MOSRS) | |
50 | ukca_solve_coagnucl failures | #319 (MOSRS) | UKCA-GLOMAP | Example fix here (MOSRS) | Mainly a workaround To be lodged | |
51 | Units of wetland emissions | #571 (MOSRS) | All versions | Example fix here (MOSRS) | Wetland emissions misinterpreted as gridbox mean rates rather than a land fraction rate - Lodged at UM10.1 as part of the UKCA bugfix_package under ticket #564 (MOSRS) | |
52 | surface area array so4_sa in ukca_main with L_ukca_sa_clim true setting so4_sa to zero if CLASSIC is off | #609 (MOSRS) | UKCA with l_het_psc | Needs further investigation. NS to look into this and find more appropriate owner. | ||
53 | UKCA RADAER namelists aren't read on the Cray | #577 (MOSRS) | RADAER, some platforms only | Change to input files, not code | lodged at UM10.1. MR to update this with his changes to make this work at earlier UM versions. Ben Johnson to discuss with UM systems/collaboration teams and ask for best approach to making available. | |
54 | Offline oxidants Chemistry: Diagnostics not properly initialised |
#659 (MOSRS) | Offline Chem | None yet | to be lodged at UM10.3 | |
55 | Use of trigonometric variables e.g. f3_at_u not consistent with ENDGame |
#658 (MOSRS) | All UKCA | vn10.2_ukca_bugfixes_package | Basic fixes included in #805 to be lodged for 10.3 | |
56 | Array size issues with UKCA tracers when using EG limited area models | #682 (MOSRS) | ENDGAME UKCA LAMs | MOSRS: vn10.1_ukca_eg_array_dims | Fixed at 10.3 | |
57 | Incorrect logic for DMS anciliaries required | #708 (MOSRS) | UKCA jobs not using CLASSIC but without DMS in chemistry |
MOSRS: vn10.1_aqum_eg_fixes | Model stops with message about missing input. NS. In progress. Very minor, only affects jobs without CLASSIC which have no DMS in mechanism. | |
58 | Fastj-X runs on patches all in dark | None yet | UKCA jobs using FastJ-X | None yet | May not be a bug, but if so line should be deleted not commented out | |
59 | Mismatch in levels while reading CDNC from D1 under ENDGame |
#704 (MOSRS) | EG-UKCA vn9.0+ with Indirect effects |
MOSRS: vn10.0_fix_cdnc_lev | To be lodged at UM10.3 in #805 | |
60 | Model failure with PrStatus_Diag | #784 (MOSRS) | UKCA on Cray | MOSRS: vn10.2_ukca_prstat_diag_fixes | To be lodged at UM10.3 in #805 | |
61 | MPI BCAST bug in NetCDF emiss code | #782 (MOSRS) | vn10.1 UKCA with Netcdf emiss |
MOSRS: vn10.2_ukca_netcdf_bcastbug | To be lodged at UM10.3 in #805 | |
62 | Inconsistency in aero_surf area used in heterogeneous reactions routine. Het rates to be placed in the UM dump file if a job is to be restarted |
(MOSRS) | UKCA+ het | To be lodged. CJ | ||
63 | Inconsistency in timestep/ interval calculation for BE-Offline scheme |
#774 (MOSRS) | All BE Offline | vn10.2_ukca_repair_interval | To be lodged at vn10.3 | |
64 | Very low cloud fraction values can cause failures in Fastj/jx |
#757 (MOSRS) | Any | Simple fix: vn10.2_ukca_acf_zero |
Simple fix lodged at vn10.3 | |
65 | qcl not correctly initialised in ukca_calc_aqueous_6a |
Part of #805 (MOSRS) | UKCA run with 6A convection |
MOSRS: vn10.2_ukca_bugfixes_package | To be lodged at UM10.3 in #805 | |
66 | Crash in NetCDF emissions and oxidants routines when using NetCDF-4 libraries |
NA | vn10.0-10.2 UKCA + NetCDF |
MOSRS: vn10.2_ukca_readnc_offline_fix | Automatically fixed as part of #282. Only affects coupled model versions. JM to follow up |
Targets for vn10.3
- MD is aiming to lodge a bug fixes branch (#805, see above) and has updated the UKCA apps to more recent versions incuding use of NetCDF emisisons
- JD is looking to lodge the aerosol tunings needed in response to the fix to the nucleation scavenging changes
- Ben Fitzpatrick has found a minor bug with an INTENT problem.
- Stephanie Woodward is working on a GLOMAP dust only configuration. The code to get a version with only insoluble dust modes on was ldoged at 10.2 and for 10.3 the aim is to support turning on coupling to radiation and perhaps indirect effects. No ageing into soluble modes is implemented at this point.
- NS and others have already lodged a number of fixes to allow an AQUM job to work with ENDGAME. Those not covered above are in CLASSIC and so not relevant here.
- NS is hoping to lodge a major change to rationalise the way non-transported prognostics (NTPs) are dealt with in the code. Jobs properly upgraded using the upgrade macros will have all reference to the old stashcodes removed and the NTPs in their new slots will be set to zero.
- NS has in the progress of working on the NTPs found an issue with time availability codes for UKCA chemistry timesteps - these have not been implemented correctly, or used with any UKCA diagnostics. NS to talk to Jane about her requirements and open ticket to capture issues. Also see if CRUM team will take on the work.
- There is also an issue of the NTP with output of NTPs on pressure levels. NS to try and resolve this before lodging the code, but if not possible open a ticket to investigate further after 10.3
- AH - one ticket with reconfiguration work allowing the use of GLOMAP-mode climatologies now in sci/tech review and should be readyf or 10.3. A second ticket will allow netCDF files to be input by the model and is hopefully going to be lodged for 10.4.
- FOC is working on a change to UKCA to respond to changes to the wetlands code so that UKCA picks up the correct emissions of methane when the basis of the scheme is changed.
LFRic
- Matt Hort (Hd, Atmospheric Dispersion and Air Quality) is seeking additional Met Office resource to work on composition in LFRIC.
- LA has spoken to John Pyle, and he will discuss resources for LFRic with Rowan Sutton.
Agreed by all that this has to be a partnership between NCAS and the Met Office and that cannot be done by one organisation on its own.
AOB
Paul Griffiths will now be looking at upgrading Fast-JX to implement the new "Cloud-J" scheme which is the latest addition to the code from Michael Prather. AA to investigate he can also consider the coupling of Fast-J to GLOMAP-mode
Date of next meeting
FOC to set up Doodle poll for 1 week in December and the second week in January