UKCA Release Jobs

This page contains documentation on UKCA standard jobs that have been released or will soon be released.

You may also wish to look at the bugfixes page.

You may also wish to look at the MONSooN IBM to Cray Transition page.

Please also see the Getting Started with UKCA page.

Standard Release Jobs

UKCA can be configured through the UMUI, via it's own configuration panel. This mostly controls all the options required to run UKCA, although some extra changes have been made with hand-edits. To make it easy to get started with UKCA, several standard jobs have been created at various resolutions. To get started with UKCA, it should be a simple matter to take a copy of the UKCA job that you wish to run, changing a few user-specific options, and submitting the job.

Details of the jobs on ARCHER can be found here.

Code Changes

To run a UKCA standard job, please take a copy of one of the jobs listed below. If you need to make code changes, you must make a new FCM branch, and merge in the UKCA branch which is used in the job that you are taking, then use your branch instead.

Please do not take a checkout of the UKCA branch used in the UMUI FCM panel (e.g. fcm:um_br/dev/luke/VN7.3_UKCA_CheM) and run from this as a working copy. If the fcm commit command is run from this directory it would commit any changes you have made to this branch. In this case, this would affect the branch for other UKCA users.

Job List

Available release jobs are:

Release Job #	UM Version	UMUI Job-ID	Climate Model	Horizontal Resolution	Vertical Levels	Model Top	Scheme Used	FCM Branch	Revision Number	Documentation
RJ1.0	7.3	HECToR:xfvfa,xfvfc MONSooN: xfvfb ARCHER: xfvfd	QESM-A	N48 (2.5°×3.75°)	L60	~84km	Tropospheric Chemistry with Isoprene	fcm:um_br/dev/luke/VN7.3_UKCA_CheM	13783	TropIsop
	7.3		HadGEM3-A r2.0	N96 (1.25°×1.875°)	L63	~40km	Tropospheric Chemistry with Isoprene
RJ2.0	7.3	HECToR:xfvga,xfvgc MONSooN:xfvgb ARCHER: xfvgd	QESM-A	N48	L60	~84km	Stratospheric Chemistry	fcm:um_br/dev/luke/VN7.3_UKCA_CheM	13783	StratChem
	7.3		HadGEM3-A r2.0	N96	L85	85km	Stratospheric Chemistry
	7.3		QESM-A	N48	L60	~84km	Simple Tropospheric Chemistry with MODE Aerosol
RJ3.0	7.3	HECToR: MONSooN:xfxlb		N96	L38	~40km	Simple Tropospheric Chemistry with MODE Aerosol	fcm:um_br/dev/mdalvi/VN7.3_MODEBLN_Rad_Indir	5953	UKCA-MODE
RJ3.1	7.3	HECToR: MONSooN:xfxld		N96	L38	~40km	Simple Tropospheric Chemistry with MODE Aerosol (Nudged)	fcm:um_br/dev/mdalvi/VN7.3_MODEBLN_Rad_Indir fcm:um_br/dev/mdalvi/VN7.3_nudged_new	5953 5586	UKCA-MODE
RJ4.0	8.4	ARCHER: xlavb MONSooN: xlava	GA4.0	N96	L85	85km	CheST chemistry with GLOMAP-mode aerosols	fcm:um_br/pkg/Config/vn8.4_UKCA	18311	CheST+GLOMAP-mode
RJ5.0	11.0+	See GA7.1 StratTrop suites table for nudged and free-running	GA7.1	N96e	L85	85km	StratTrop (CheST) with GLOMAP-mode aerosol	Runs from the trunk with possible additional branches		Release Job UM11.0

Required UMUI Changes

Before submitting a UKCA job, you must first alter several options in the UMUI:

• User-id, Mail-id (email address), and Tic-Code:
  Model Selection -> User Information and Target Machine -> General Details
  The user-id is your username on the target machine (i.e. HECToR).
  The tic-code will be the HECToR charging code, e.g. n02-chem. This code can be found by logging into the HECToR SAFE pages, and viewing your account information.
• Target machine root extract directory (UM_ROUTDIR):
  Model Selection -> FCM Configuration -> FCM Extract and Build directories and Output levels
  This needs to be the full path to one of your directories on the target machine (HECToR/ARCHER /home directory or MONSooN /projects directory).

See also: Using the UMUI for information on Optional UMUI changes such as changing the run length, archiving, emissions etc

Tropospheric Chemistry with Isoprene - TropIsop

See also Tropospheric Chemistry with Isoprene description page.

Comparison of Tropospheric Chemistry with Isoprene in a number of model resolutions and UM versions.

This scheme using the FCM branch fcm:um_br/dev/luke/VN7.3_UKCA_CheM at revision 13783 (updated). It has 56 species (49 of which are tracers), 115 bimolecular reactions, 35 photolysis reactions (using rates determined using a 2D look-up table) and 15 termolecular reactions. 23 species are wet-deposited and 32 are dry-deposited (using a 2D dry-deposition scheme). Tropospheric $O_{x}$ and $CO$ budgets are outputted as standard, as are other diagnostics of interest, such as the $CH_{4}$ lifetime and Stratosphere-Troposphere Exchange of $O_{3}$.

The chemistry is integrated using the ASAD chemical package, using a sparse-matrix Newton-Raphson solver. While the UM dynamical timestep is every 20 model minutes, chemistry is called every model hour, although emissions are still performed every dynamical timestep. Top boundary conditions are used for $O_{3}$ and $NO_{y}$ and $CH_{4}$ is treated as a constant. The tracers are initialised to spun-up fields from a previous TropIsop run, however, we recommend that a 15 month-spin-up should be performed.

The chemistry scheme used in these simulations is the same as in Telford et al, Atmos. Chem. Phys., 10, 7117-7125, doi:10.5194/acp-10-7117-2010 (2010). A model description paper is also in preparation (O'Connor et al, 2011).

The tropospheric ozone concentrations in UKCA-TropIsop are biased slightly high. This is improved with using an interactive photolysis scheme, which will be released for general use soon.

Timings

It should be noted that due to "jitter" on the HPC platform, all timings are approximate. As a simulation progresses the time per model month may increase as the fields spin-up from their initial state. Please also note that the MONSooN timings are for phase 1 - jobs will run more quickly on phase 2 and you should consider using more nodes than on phase 1.

Machine	No. of Cores	No. of Nodes	Resolution	Time for 1 Model Month
HECToR Phase2b	96	4	N48L60	~50 minutes
MONSooN Pahse 1	64	2	N48L60	~59 minutes
ARCHER	72	3	N48L60	~27 minutes

Stratospheric Chemistry - StratChem

See also Stratospheric Chemistry description page.

Ozone column in Dobson Units (DU) calculated from a UM7.3 N48L60 UKCA-StratChem run using interactive UKCA ozone in the UM radiation scheme. This is the mean of 21 model years representing 1990.

This scheme using the FCM branch fcm:um_br/dev/luke/VN7.3_UKCA_CheM at revision 13783 (updated). It has 41 species (36 of which are tracers), 113 bimolecular reactions, 34 photolysis reactions (using rates determined using a 3D look-up table), 17 termolecular reactions and 5 heterogeneous reactions. 15 species are wet-deposited and 15 are dry-deposited (using a 2D dry-deposition scheme). A stratospheric $O_{3}$ is standard (based on the budget of Lee et al, JGR 107 DOI:10.1029/2001JD000538 (2002)), as are other diagnostics of interest, such as the $CH_{4}$ lifetime and Stratosphere-Troposphere Exchange of $O_{3}$.

The chemistry is integrated using the ASAD chemical package, using a sparse-matrix Newton-Raphson solver. While the UM dynamical timestep is every 20 model minutes, chemistry is called every model hour, although emissions are still performed every dynamical timestep. By default, lower boundary conditions set for the tracer concentrations of $N_{2}O$, $CH_{4}$, $H_{2}$, $CF_{2}Cl_{2}$, $CFCl_{3}$, $CH_{3}Br$, and $CH_{2}Br_{2}$ are specified for the years 1950-2100. These values are also used in the UM radiation scheme, along with $CO_{2}$ concentrations. The tracers are initialised to spun-up fields from a previous StratChem run, however it can take time for a Stratospheric run to equilibrate and spin-up times as long as 20 model years may be required if significant changes are made.

The chemistry scheme used in these simulations is the similar to that described in Morgenstern et al, Geosci. Model Dev., 2, 43-57, (2009).

Please also see the bugfixes page for further information on the initial conditions used in the release job.

Timings

It should be noted that due to "jitter" on the HPC platform, all timings are approximate. As a simulation progresses the time per model month may increase as the fields spin-up from their initial state. Please also note that the MONSooN timings are for phase 1 - jobs will run more quickly on phase 2 and you should consider using more nodes than on phase 1.

Machine	No. of Cores	No. of Nodes	Resolution	Time for 1 Model Month
HECToR Phase2b	96	4	N48L60	~46 minutes
MONSooN Phase 1	64	2	N48L60	~51 minutes
ARCHER	72	3	N48L60	~25 minutes

Simple Tropospheric Chemistry with GLOMAP-mode aerosol - UKCA-MODE

See also MODE Aerosol description page.

The UKCA aerosol scheme 'GLOMAP-mode' (sometimes called UKCA-MODE when in the UM) is described in detail in Mann et al (2010) and simulates aerosol microphysics (nucleation, condensation,coagulation and cloud processing), dry deposition, sedimentation and in-cloud/below-cloud scavenging. The scheme simulates atmospheric aerosol in a size-resolved manner with the chemical and aerosol microphysical processes determining the evolution of the size and composition of several size modes. The scheme transports the mass of sulphate, sea-salt, black carbon, organic matter and dust, along with the particle number concentration in each size mode. The aerosol scheme is coded flexibly to allow the same code to run with several alternative aerosol configurations via FORTRAN-90 modules.

UKCA-MODE can be coupled to all of the various UKCA chemistry schemes, but in this standard job uses the simple tropospheric chemistry scheme, extended to treat the degradation of the aerosol precursor species: sulphur dioxide (SO2);dimethylsulphide (DMS); and a lumped monoterpene tracer.

An interface between UKCA-MODE and the Edwards-Slingo radiation scheme allows the aerosol concentrations to interact with the radiation both directly (through scattering and absorption of incoming solar and outgoing long-wave radiation) and indirectly via modified cloud albedo. In the standard jobs, both these effects are computed in a passive/ diagnostic mode via use of a double-call to the radiation scheme.

One of the standard jobs also includes Nudging, wherein the model dynamics variables are nudged or relaxed towards ECMWF Reanalysis data. This configuration is particularly useful for evaluating the MODE scheme against measurements.

Timings

When run with all the 31 aerosol tracers switched on, the UKCA-MODE code adds 12% to the cost of the Atmosphere model (see O'Connor et al, 2011 ICP report to DECC/DEFRA). Please also note that the MONSooN timings are for phase 1 - jobs will run more quickly on phase 2 and you should consider using more nodes than on phase 1.

Machine	No. of Cores	No. of Nodes	Resolution	Time for 1 Model Month
MONSooN (non-nudged)	64	2	N96L38	~2 hours
MONSooN (nudged)	64	2	N96L38	~2.5 hours

Useful Links

• Restarting an Atmos-Ocean Integration
• Adding New UKCA Tracers
• Duplicating STASH between UM jobs
• Using the UMUI
• NCAS-CMS
• Subversion (SVN)
• Flexible Configuration Management (FCM)
• UM FCM Tutorial
• HECToR
• Xconv
• Further Xconv help
• Xancil
• ASAD