Difference between revisions of "GA7.1 StratTrop suites"

From UKCA
Line 37: Line 37:
 
|| '''UM11.9''' || ''<code>u-cd037#</code>'' <br/>'''<code>u-cd694</code>'''|| || ||
 
|| '''UM11.9''' || ''<code>u-cd037#</code>'' <br/>'''<code>u-cd694</code>'''|| || ||
 
|-
 
|-
|| '''UM11.9 GA8.0''' || '''<code>u-cf821</code>''' <br/> {{pdf|ukca_eval_cd037_cf821.pdf|Comparison with GA7.1}} || || ||
+
|| '''UM11.9 GA8.0''' || '''<code>u-cf821</code>''' <br/> {{pdf|Ukca_eval_cd037_cf821.pdf|Comparison with GA7.1}} || || ||
 
|-
 
|-
 
|}
 
|}

Revision as of 15:50, 19 July 2021

This page contains details of GA7.1/GA8.0+StratTrop suites available.

Usage

- COPY these suites for your use rosie copy u-nnXXX (or in Rosie GUI), rather than checking them out.This will prevent you accidentally over-writing the settings in the original suite.
- For Monsoon2, remember to change the project-name before running (This might be set to ukca-meto by default), otherwise you are likely to see submit-failed error messages for tasks running on the HPC.

TS2000 free-running suites

These are year-2000 atmosphere-only timeslice simulations.

UM versions Met Office XC40 Monsoon2 XCS NEXCS ARCHER
UM11.0 u-ay770 u-bb055 u-bb056 u-ba915
UM11.1 u-bd086 Will be produced on request
UM11.2 u-bd036 u-bd683 u-be045
UM11.3
Bit-reprod with UM11.2
u-bg745 u-bg779 u-bg785
UM11.4
Pdficon small.png Comparison with UM11.3 Info circle.png
u-bk249 u-bl029
UM11.5
Pdficon small.png Comparison with UM11.4 Info circle.png
u-bo717 u-bo741
UM11.6
Pdficon small.png Comparison with UM11.5 Info circle.png
u-bs395 u-bs425
UM11.7
Pdficon small.png Comparison with UM11.6 Info circle.png
u-bw489 u-bw692
UM11.8 u-bz814 u-bz918
UM11.9 u-cd037#
u-cd694
UM11.9 GA8.0 u-cf821
Pdficon small.png Comparison with GA7.1 Info circle.png

#: u-cd037 has an issue with fields requested with sub-daily sampling (e.g. T6HDPMN) where all the daily values are written to the monthly mean files. Does not affect fields with TMONMN, but the pm files are bigger in size. Use of u-cd694 is recommended, which has the correct settings.

  • An example vn11.2 ARCHER suite configured for a 1-day run without archiving and with manual compile is u-be079. A selected set of 3-hourly meaned UKCA diagnostics are also sent to the daily UPK stream.

TS2000 nudged suites

These are year-2000 atmosphere-only timeslice simulations nudged to ERA-Interim reanalysis. Here the nudging files will change with date rather than being a perpetual year, although the forcings are all climatologies.

Instructions for setting up an ERA dataset for perpetual year Nudging can be found here: Nudging perpetual

UM versions Met Office XC40 Monsoon2 XCS NEXCS ARCHER
UM11.0 u-ba749 u-bb202 u-bb210
UM11.1 u-bd267 Will be produced on request
UM11.2 u-bd359 u-bd341 u-bd366
UM11.3 u-bh303
UM11.4 Will be produced if needed
UM11.5 u-bq095
UM11.6 u-bt030
UM11.7/UM11.8
UM11.9
using ERA5
u-cd061 u-cd679

Release notes / Known Issues

All UM versions (vn10.6.1+)

  • Stash items spanning 38:485 to 38:545 i.e. Aerosol surface conc., number conc., load diagnostics: Due to the way STASH is handling and processing these items -when requested using Stash time profiles like TDMPMN, TMONMN, TDAYMN- will be output as exactly 1/3rd of their actual value. Members making use of these items need to scale the values by 3.0.

All UKCA configurations till UM11.2 using ENDGame

  • The Strat-trop exchange diagnostics (e.g. O3 STE 50-051) are incorrect due to errors in handling of the extra model-level (introduced under ENDGame dynamical core) during their calculation.

All UKCA configurations till UM11.6

  • The O(1D) mass mixing ratio field (STASH 34997/ 51997) is incorrect due to an error in the conversion factors used in its calculation.

Testing suites

Further details on UKCA testing suites on Monsoon2 can be found at

Other Configurations and Input Files

Please see these pages for details of further UKCA configurations in use, and information about other input files:

  • Other UKCA Configurations
  • Other UKCA Input Files
  • CMIP6 Emissions: Details of the procedure used for generation of emissions and other forcing data for UKESM1/CMIP6 expts can be found in links from this page. Wherever possible, first try to contact the creator/ reviewer of the particular dataset in case you need more information

Converting Y2000 suites to transient

Instructions for converting the above Y2000 (decadal mean forcing) suites to transient can be found here.

Virtual Machine Development Suites

Suites that can be run on the Virtual Machine are listed here. It is relatively straight-forward to upgrade suites to required versions due to the simple nature of the suite.

  • All jobs are column-call configurations except for UM10.9/u-as159. This means that you can change the decomposition and the results will stay the same.
  • All jobs run for 3 model hours
  • The N48L38 jobs run for 6 timesteps (30 min timestep) and needs around 4GB of memory free to run (a minimum of a 6GB VM is required)
  • The N48L70 jobs run for 9 timesteps (20 min timestep) and needs around 8GB of memory free to run (a minimum of a 10GB VM is required)

For further information see Abraham et al. (2018).

UM versions N48L38 N48L70
UM10.9 u-as159
UM11.0 u-ax117
UM11.1 u-bb521 u-bc444
UM11.3 u-bh045
UM11.5 u-bo389 u-bo378
UM11.7 u-by854 or u-bz223 u-by855
UM11.8 u-ca021
UM11.9
UM12.0 u-cg192 (no prebuilds)
u-cg311 (with prebuilds)