#############################################################################

RUNNING postprocessing

###########################################################################

add the following row to file mip_info.csv

MIPNAME - identifier used in code to select the relevant information from mip_info.csv

- include 'CMIP' or 'TRENDY' or 'ISIMIP' in this name depending on what the MIP is.

- you might need to add for another mip, but try and use one of these if possible

- the differences between them are mainly in the syntax of output file

- 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_name will 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 file name (e.g. land-hist, historical, rcp2p6, c20c)

in_scenario - scenario for input file name (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 files

end_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] MIPNAME

l_testing switch: test runs for the first 10 years of data for the variables in sel_diags_test.

l_backfill_missing_files: this determines whether the file for that variable exists already if it exists it doesnt make a new file

l_jules_rose: this is used when need the postprocessing 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

The variables to be postprocessed are defined in ***_variables.py where *** is the first term (before the underscore) of the MIPNAME. This is only used outside rose.

#############################################################################

ISSUES

################################################################### CMIP6: fNdep: should be monthly (Emon) and not annual - needs to be changed in the output profiles and in the variable namelists

################################################################### ISIMIP: 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.

#############################################################################

RUNNING MIP_CONVERT to make CMIP data for submission

########################################################################### Getting hold of the code; fcm co https://code.metoffice.gov.uk/svn/cdds/main/branches/dev/matthewmizielinski/r8439_1898_JULES_land_only

Activate code; cd to r8439_1898_JULES_land_only run “source ./setup_env_for_devel”

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 lnm check for CRITICAL messages in the output files (mip_convert_.log)

Run mip_convert for dailies mip_convert mip_convert.cfg -s lnd check for CRITICAL messages in the output files (mip_convert_.log)

Check the files look OK. (last chance to review)