Minutes of the code management group meeting 2026-01-08

Code management group minutes

Teleconference Details

Thursday 8th January 2026, 10.00am - 11.30am GMT

Meeting via Teams - see link in calendar invitation

Met Office: MD, AJH, Pedro, MS

Cambridge: RW

Leeds: MR

Canterbury: ??

CSIRO: Ashok

Apologies: NLA , DG

NEXT MEETING WILL BE THURSDAY 12TH March 2026, 10.00-11.30 GMT

Actions from this meeting

AJH to request UKCA tags for lfric_apps and UM , so that we are able to keep track of all UKCA GitHub pull requests and issues at the next CMG meeting.
MD to put a link to Monsoon3 guide to GitHub in the minutes.
MD is developing workaround for Bug in Lfric copy of ukca_volcanic_so2. However, this will not add functionality in lfric. A developer for this needs to be identified.
AJH has contacted the LFRic core team to inform them of the obsolete instructions for building lfric core with FCM.
AJH to tidy up the priority table. Some items are now complete and should be marked as such.
AJH and Pedro to discuss method for checking diagnostic KGOs with NetCDF files.

Actions from the Last Meeting

(ongoing) GM/DG to update table on UKCA wiki documenting GLOMAP versions. DG/GM to consider the code and lodging process for GLOMAP developments and consider a suitable rose-stem test supported by NLA. GM Will work with DG to get a subset to commit. AJH to review. Graham has found code needed for meteoric smoke.
(ongoing) AJH on update API document to UM13.6 release. NLA to provide feedback before it is be circulated. There is a comprehensive overhaul for 13.7. AJH to make document available for wider information/comment - see here. AJH won't do anything with API until GitHub release is out, and then will add some things to it.
(ongoing) ALL look at um:#6203 showing other unprotected rose-stem options and think about any code/settings they are interested in. May be able to use GitHub actions.
(ongoing) ALL to consider ordering of the Progress with Priorities table and highlight things that should be higher priority, and any other tasks that need to be included.
(complete) AJH to chair next meeting
(this happened - not sure if follow up is required) NLA to mention MASS issues at the next MMG
(ongoing) NLA to put a link to Monsoon3 guide to GitHub in the minutes

Tickets

This is a work in progress. AJH and MD to figure out how UKCA keywords work in Github

AJH is developing this change https://github.com/MetOffice/lfric_apps/pull/94

Mohit is developing this change https://github.com/MetOffice/lfric_apps/issues/81

Rob Waters is targeting the current release to get completed FCM ticket into GitHub

Mark R is continuing an NG-Arch development - but this probably wont make the 30th January cut-off.

Code submission deadline is Friday 30th January.

Quick tour Github

AJH - sorry I was unable to record the meeting as i didnt have permissions.

I gave a brief tour of the development of my bug fix using the Github system and brought up some current teething problems with the new system.

It is pretty easy to scrap something in Github and start again, and I made almost every possible mistake when working on https://github.com/MetOffice/lfric_apps/pull/94

I made a note of the Git commands (aka my cheat sheet) that I used here. Note that my lazy method of merging my FCM changes (using xxdiff) worked, but is not the recommended method. https://github.com/MetOffice/lfric_apps/issues/41#issuecomment-3660674635

I would recommend developers start with an easy ticket for their first Github development.

Should I continue with MOSRS or move to Github

Ashok - Should I continue with MOSRS or move to Github AJH - If your development is almost complete, it might still be useful to finish up in MOSRS first.

It is probably best to move stright to Github for almost all current developments.

MOSRS tickets and branches can still be modified, but this functionality will be removed at some point in 2026.

What to do with MOSRS pages

RW - Do I need to close my MOSRS pages AJH - MOSRS is essentially dead to the UM systems team. You are welcome to tidy them up or to do nothing with them.

Later only the committed tickets will be permanently preserved. Do not modify these.

Last Release

Progress with Priorities (NLA)

ACTION REQUEST FROM UKCA SCIENCE & MANAGEMENT BOARD: CMG to provide list of priorities & their status, with particular thought to UKESM2 changes that will be required.

Progress with Priorities table

For diagnostics it was decided to have 2 levels

~~Diagnostic infrastructure and high priority diagnostics that are needed for auto-assess and the UKCA evaluation suite (new #7)~~
All other diagnostics (new #11 after GPU work)

For people who are using LFRic, only imode_setup=8 works for RADAER. AJH is aiming to get back to expanding this list of options in December.

Bugs/Bugfixes (NLA)

Open Bugs

#	Description	Ticket	Owner	Comments	Comments from this meeting
124	H2 and N2 cannot be set for RCP and WMOA1 scenario routines	UM:#5620 vn13.8.1_ukca_rcp_read_h2_n2	Mohit	Values are hard-wired in the code and cannot be changed without a branch. N2 values seem to be read from namelist (for UM, and RCP scenarios). Branch adds possibility to read H2 from either RCP or namelist. Will probably require updating all UKCA-chem apps to maintain KGO. More testing required MD has done a little bit of work. N2 is read from namelist, but can now read from RCP file for H2
128	Bug in UKCA stratospheric chemistry diagnostics	UM:#5839	Luke Abraham	The indexing/values of diagnostics m01s50i101-i104 are incorrect. Fixes provided for vn11.1 and vn11.7. May affect CMIP6 diagnostics Does this need fixing for CMIP7? Need to discuss with CDDS/diagnostics team as might break existing workflow.
143	Error in print statements from UKCA_TRACERS_COPY_FROM_UM	Not yet	???	The final species listed in the diagnostic print statement seems to have very large concentrations (O(1E+08)) compared to others in the ppm-ppt range. Does not seem to affect model output. Reported by James Weber.
144	interchangeable use of specific humidity and water vapour mixing ratio	#6813	John Hemmings and ???	UKCA includes calculations of relative humidity, clear-sky relative humidity and saturation vapour pressure. Relative humidity calculations should be based on water vapour mass mixing ratio but are in part based on specific humidity. Also, the calculation of saturation vapour pressure is derived from saturation mixing ratio but uses a formula for deriving SVP from saturation specific humidity. This interchangeable use of specific humidity (ratio for moist air) and mixing ratio (ratio for dry air) in existing configurations introduces small numerical errors in UM and LFRic model runs. These bugs were identified at UM version vn12.2 and have been highlighted by comments in code lodged with UM ticket #6660 (in ukca_main1-ukca_main1.F90 and atmos_ukca_humidity_mod.F90). All UKCA configurations are affected. Alan - This should be fixed by Met Office. I will speak with Anthony and Alistair about resourcing to fix this. It requires good understanding of humidity to work through this problem, otherwise I would have done this myself. When we pass stuff through the argument list, the name changes which makes things complicated. Should we pass a structure with all the humidity fields in? There is a UKCA bespoke version of humidity calculations for non-UM applications. Internal Met Office problem - will need to find someone to fix it. AJH to talk with AJ and Alistair Sellar. GM can cross-check. Ticket has 2 parts. JH suggests doing it in 2 tickets, 1 technical, 1 bug fix. JH could do technical part (either as callback routines or separate fields). Then would need UM & LFRic tickets for the bug fix, as calculations could be done in the parent or as callback routine. Some work has been done - technical change has been done in ukca:#245. Went in at UM13.9. This option is now used by the UM and LFRic. Buggy calculations are now outside of UKCA - will require UM and LFRic changes to fix those. Might not be a large amount of work, but problem needs to be well understood. Also consider what field is being passed in the tracer list - currently q but should be mixing ratio. Some has been fixed - bug has been shifted outside of UKCA. Don't know of anyone working on it currently. Someone will need to go through and check we are passing the correct one - need someone on aerosol side and on chemistry side
153	bug in 2D photolysis			While working on developments for the new photolysis API, we’ve noticed a likely error in the current implementation of the 2D photolysis. The 2D photolysis files are tabulated by day of year at 5-day resolution, amongst other variables. The interpolation to the actual day of year is currently done in the read subroutine (read2d_opt or read2d_orig). However, this subroutine is only called once on the first timestep. For multi-day runs, the photolysis rates will not be updated as the day of year increases. Since nobody is using the 2D photolysis at the moment this is likely of low to no impact. Might be used by NUAQ BD or VB will know more - PMJ will ask
154	bug in tropospheric heterogeneous chemistry			Within GLOMAP the `wetdp` variable is defined as the wet diameter in m. However, in `ukca_trop_hetchem` this is used as a radius in cm. This results in a factor of 50x error. Steve Turnock might be using this. NLA to contact ST and YG about this.
155	OMP dynamic scheduling race condition	ukca:#321		When using OpenMP with dynamic, as opposed to the standard static, scheduling, checksums were failing with lfric_apps. This indicates a race condition likely within the ASAD chemical package and so would affect the UM as well as LFRic. Further testing is needed.
156	Bug in Lfric copy of ukca_volcanic_so2	Github lfric_apps issue 81	Mohit	Fix being devloped	Mohit is targetting the current release. Its quite an urgent fix, as it breaks lfric on some platforms
157	Bug in calculation of aerosol AOD from LFRic	Github lfric_apps PR 94	Alan	Fix in review	Under review

Round (virtual) Table of current activities, plans

Person	Activity
AJH	I now have a plan for adding different i_mode_setup for RADAER into LFRic_apps
Mark R	Is continuing from Jon Elsey developments in NG-ARCH
Pedro	Finish the inclusion of the PM diagnostics in John Hemmings new UKCA diagnostic handling system. This change impacts UM too, as any configuration model which request PM diagnostics need to use the new system. Modify the AQUM workflow to retrieve and compile UKCA/UM code from GitHub. Investigate the status of the rose app-upgrade macro to easily allow me to upgrade the AQUM workflow. This will allow me to upgrade the workflow to the latest version of UM/AQUM/Jules, etc from the latest versions hosted in Trac to the new version hosted in GitHub.

A.O.B. & D.N.M. (NLA)

AJH - LFRic diagnostics are not currently being tested for KGO and could potentially get broken. there are no plans to make this system work, so I am gathering interested parties to come up with a solution to protect diagnostics.

Pedro - In name we have already a method for checking NetCDF KGO changes that could be useful for UKCA.