Add the following row to mip_info.csv:
MIPNAME: identifier used in code to select relevant information frommip_info.csv- include
CMIP,TRENDY, orISIMIPin this name depending on the MIP - you might need to add for another MIP, but try and use one of these if possible
- differences are mainly in output filename syntax, plus small formatting details and sometimes variable names
model: model name for output file (e.g.JULES-ES,UKESM1-0-LL,HadGEM3-GC31-LL)suite_id: usually rose suite name (src_dir+suite_name = path to data; this is not always rose suite id)src_dir: base path to input data (suite_namewill be added to find actual data)run_name: start of filename for input data (as defined in JULES output profiles)out_scenario: scenario for output filename (e.g.land-hist,historical,rcp2p6,c20c)in_scenario: scenario for input filename (e.g.S2,S3)start_year_in: start year for output (if not all available files are used); set to less than 0 to read all available filesend_year_in: end year for output (if not all available files are used); set to less than 0 to read all available files
On the Met Office Linux system:
module load scitools
python process_jules.py [-h] [--l_testing] [--l_backfill_missing_files] [--l_jules_rose] MIPNAMESwitches:
l_testing: test runs for the first 10 years of data for variables insel_diags_testl_backfill_missing_files: checks whether the file for that variable exists already; if it exists it does not make a new filel_jules_rose: used when postprocessing needs to run within the JULES rose suite
MIPNAME is used to look up information in mip_info, which gives details of the model to be postprocessed. This is used outside rose.
Variables to be postprocessed are defined in ***_variables.py, where *** is the first term (before the underscore) of MIPNAME. This is only used outside rose.
fNdepshould be monthly (Emon) and not annual- this needs to be changed in the output profiles and in the variable namelists
- need to get final year of future simulation
- this can be done by adding another year to the driving by making links to the last year
Getting hold of the code:
fcm co https://code.metoffice.gov.uk/svn/cdds/main/branches/dev/matthewmizielinski/r8439_1898_JULES_land_onlyActivate code:
cd r8439_1898_JULES_land_only
source ./setup_env_for_develDirectory layout:
working_directory/
input/
<suite_id>/
lnd/ (daily data files)
lnm/ (monthly data files)
output/
mip_convert.cfg
Get hold of config file (download from):
https://code.metoffice.gov.uk/trac/cdds/ticket/1898
Adapt as needed: change experiment_id, variant_label, suite_id, run_bounds, etc.
Run mip_convert for monthlies:
../r8439_1898_JULES_land_only/mip_convert/bin/mip_convert mip_convert.cfg -s lnmCheck for CRITICAL messages in output files (mip_convert_<datestamp>.log).
Run mip_convert for dailies:
mip_convert mip_convert.cfg -s lndCheck for CRITICAL messages in output files (mip_convert_<datestamp>.log).
Check files look OK (last chance to review).