GEMCLIM config files

There are three configuration files used when running GEMCLIM.

Two of these are read by the executables (entry and/or model):
The third one is only used in batch mode:


gemclim_settings.nml

This file contains all the namelists to control the grid and the different schemes and parameters for the entry program (gemclimntr) and the main program (gemclimdm). The namelists
Information about the variables in the grid, entry and config namelists can be found here and the information about the physics namelist is here.

Special variables only available (in version 3.3.2)  in GEMCLIM:
namelist 'gem_cfgs':
Out3_pilstep
 If 'Out3_pilstep > 0' time step interval at which to output pilot files.
Default: Out3_pilstep = 0
Out3_anal_S Dates at which to write an analysis file can be specified in 'Out3_anal_S'.
Dates can be specified in form of a list of dates and/or intervals.
  Format of dates:      YYYYMMDDhh
  Format of intervals: [startdate,enddate,interval]
  Format of interval:   Increment followed by unit
  Possible units for interval: h,H: hours; d,D: days; m,M: months
Blancks are ignored.
Examples:
    Out3_anal_S = '1978030100, 1978060100'
    Out3_anal_S = '[1978013000,1978013100,6h]'
    Out3_anal_S = '[1978 03 01 00, 2008 06 01 00, 3m]'
    Out3_anal_S = '1978030100,[1978013000,1978013100,6h],1979030100'
Default: Out3_anal_S = ' '
Lam_SP_nudging_S
 Spectral nudging list of variables (eg. 'UVT').
Default: Lam_SP_nudging_S = ' '
Lam_trunc_scale
 Truncate scale for spectral nudging (km).
Default: Lam_trunc_scale = 1000.
Out3_CCCout_L
 .true. to keep variables that are the same for CCCma and RPN in MKS units.
Default:  Out3_CCCout_L  = .false.
Out3_rotate_L
 .true. to rotate vector component pairs.
Better set to .true.!
Default: Out3_rotate_L = .false.
Out3_2DLev
 real (decoded) vertical level value for 2D fields. Supported values are 0.0 and 1.0.
Default: Out3_2DLev = 0.0


The following variables should always be set according the values in this list:
namelist 'gem_cfgs':
Init_balgm_L = .false. No initialization, as it is useless in climate mode.
Schm_hdlast_L = .true. (Optional/Used for AMIP2)  Horizontal diffusion after the physics.
Schm_psadj_L = ...
 If set to .true. surface pressure adjustment is activated. This means that at every time step, the horizontal integral of pi' at the surface is restored to its previous time step value. In hydrostatic mode, ln(p) and ln(pi_s/z_s) are also corrected.
Global (uniform, variable): Schm_psadj_L = .true.
LAM mode: Schm_psadj_L = .false.
Out3_ndigits = 8 Minimum of digits used to represent output units. Required by the diagnostics scripts.
Out3_unit_S = 'P' Output names will include current steps rather than current dates and will end with a 'p'. Required by the diagnostics.
Out3_nbitg = 16 Packing factor used for all variables except for those defined in Out_xnbits_S. Please do not use anything smaller as it causes irreparable degradation of the diagnostics.
Only for debugging it is recommended to 32 bits.
Out3_rotate_L = .true.
 Rotate vector component pairs.
Clim_climat_L = .true. You will run in climate mode.
Clim_ininc_L = .true. To define physics surface forcing increments at the start of an experiment and at 0Z every 15th day of each month following that. These increments are applied by the physics package every day at 12Z in climate mode to simulate the annual cycle evolution of certain variables. Examples of affected variables are the sea surface temperatures and sea ice percentage coverage.
namelist 'physics_cfgs':
MOYHR: Must NOT be zero!


Most variables in these namelists have to be set manually. Only a few get set automatically. They are:

outcfg.out

This file controls the RPN standard file output from the GEMCLIMDM run such as frequency of output, which fields, at what levels, etc. It gets only read by the model.

You will find information about the structure of this file here.

For output intervals the first (not for the first job!) and last output hour resp. time step of 'steps' will always be updated.

A list of all model variables available for output can be found at the beginning of each model listing.
Note that this list varies depending on the schemes used!

This physics fields link shows an example of available fields for output when using ISBA.

And here is a list of the most common dynamic fields available for output:

TT
 Air temperature
 C
UU
 u-component of the wind (along the grid x-axis)
 knots
VV
 v-component of the wind (along the grid y-axis) knots
UV
 Wind modulus
 knots
GZ
 Geopotential height
 dam
HU
 Specific humidity
 kg/kg
HR
 Relative humidity
 %
ES
 Dew point depression
 C
P0
 Surface pressure
 hPa
PN
 Sea level pressure
 hPa
ME
 Mountain height
 m
LA
 Geographical latitude
 deg
LO
 Geographical longitude
 deg



configexp.dot.cfg

This file is only used in batch mode. It controls where to send the batch run, how much cpu time allowed, which machine to run the job, which period to run, which absolutes to use, which input files to use, where to place the model output, etc.

You will find information about the general variables in this file here.

Two of its variables do get updated each job with the current date:

And here is the list of special climate variables in GEMCLIM:

The variables written in red have to be initialized to run in climate mode, variables in violet only have to be set when diagnostics are required.


climat=1; Required for climate-mode GEM/GEMCLIM. Default: climat=0
UM_EXEC_exp: Experiment name. The current year and month are added/updated automatically at the start of each new job, depending also on the value of the variable ${CLIMAT_interval}: The result is thus expname_YYYYMM resp. expname_YYYYMM-MM.
Note that if ${UM_EXEC_exp} contains an underscore everything after the last underscore will be replaced by the current date.
UM_EXEC_xfer:
 Note! In climate-mode this is the location, including the machine name and directory path, where the post-processing and diagnostics will be done, also known as ${dest_mach}. Format: xfer='machine:directory'
CLIMAT_outrep:
 Set this variable to a directory on ${BACKEND_mach} which must NOT be under your ${EXECDIR}.  Because if you continue your run the old ${EXECDIR} will be erased after the next job has been launched.
UM_EXEC_ovbin:
 Location of executables. Must be on launching machine.
BACKEND_mach:
 Name of the machine where the model is to be run.
CLIMAT_startdate: Day and time when the run starts. Format:
  CLIMAT_startdate="YYYY MM DD HH";'
This date must match the validity date of the u-wind in the analysis file.
CLIMAT_enddate: Day and time when the run should finish. Format:
  CLIMAT_enddate="YYYY MM DD HH";
CLIMAT_interval: Duration of one model job in months. Every CLIMAT_interval, a new model job will be started with a call to Um_launch. There are usually several jobs in a particular experimental run. On 'marvin' keep this value set to '1'. Default: CLIMAT_interval=1
CLIMAT_rsti: Days per model subjobs (clones). Since the time a model subjob can run is limited, more than one might have to be submitted per interval. Step_rsti in gemclim_settings.nml will be set automatically according to this value. If CLIMAT_rsti is not specified, Step_rsti will not be changed.
CLIMAT_cleanup: Transfer interval in units of days. The default outcfg.out asks for a large volume of output files which may cause quota problems. To avoid this you can start the post-processing / file transfer before the model subjob is finished. CLIMAT_cleanup is the number of days after which the post-processing will be started. Step_cleanup in gemclim_settings.nml will be set automatically according to this value. If  CLIMAT_cleanup is not specified, Step_cleanup will not be changed.
CLIMAT_job_size: Sets the cpu time for several post-processing jobs. Possible values: small (default), i.e. 240x120 AMIP2 uniform grid; medium: i.e. 330x250 SGMIP grid; big: i.e. 720x360 mesoglobal grid.
CLIMAT_step_total: Number of time steps per model subjob. Calculated automatically.
CLIMAT_deltat: Size of the time step in seconds. Must be the same value as Cstv_dt_8 in gemclim_settings.nml, but is an integer.
CLIMAT_etaname: This variable sets the models vertical configuration: 1) The hybrid level distribution, 2) the pressure of the top model level. It also determines 3) the output pressure levels in outcfg.out and 4) the levels for which radiation is done (i.e. P_rad_nivl in gemclim_settings.nml). All of these etaname variables are set by the script Climat_eta2preshyb, i.e. with etaname=lh53t10; you define 53 hybrid levels with a top at 10 hPa. At the moment the following options are defined: lh28t10, lh40t2, lh42t1, lh50t5, lh50t5b, lh53t10, lh53t10b, lh60t2, lh60t2b, lh70tp1, lh80tp1, lh58tp10, lh56t10, lh64t2, lh64t2b. The default is lh40t2. To use a new vertical configuration, simply add it to your copy of the script Climat_eta2preshyb.
Usually we use:
  Global (uniform, variable): lh64t2
  LAM: lh56t10
CLIMAT_diagnos: When CLIMAT_diagnos=1, standard diagnostics will be performed, monthly means and variances.
CLIMAT_makeTS2d:
 When CLIMAT_makeTS2d=1 and CLIMAT_diagnos=1, 2D time series of certain fields will get created as part of the diagnostics.
CLIMAT_pp_cpus:
 Number of cpu's used for post processing and diagnostics. The maximum number of cpu's which can be used depends on the machine:
8 on AIX in Dorval
4 on marvin
8 on colosse
1 on most workstations
CLIMAT_fularch, CLIMAT_arch_mach,
CLIMAT_archdir:  		When CLIMAT_fularch=oui the absolutes and the restart file of each job will be archived on ${CLIMAT_arch_mach} (default 'pollux') in the directory ${CLIMAT_archdir} (default '.').
CLIMAT_save_restart:
 When CLIMAT_save_restart=1 the restart files will be archived on ${CLIMAT_arch_mach} in the directory ${CLIMAT_archdir} even if ${CLIMAT_diagnos}=0.
CLIMAT_clean: When CLIMAT_clean=1, original files will be erased after having been archived.
CLIMAT_gaussout: When CLIMAT_gaussout != 0, defines the number of grids points in the x- and y-direction for the interpolation of any global grid to a gaussian grid. The first three decimals account for the x-  and the last three decimals for the y-direction. I.e. if gaussout="240120"; the grid will be interpolated to a gaussian 240x120 grid.
The default value is 0, but setting gaussout=1 requests interpolation to 192x96 gaussian grids.
CLIMAT_splitout: Only used with global variable resolution grids. In that case, when CLIMAT_splitout=1, the high-resolution core area will be saved in files with the suffix '_hi'. Vector components will be rotated in that area. Meanwhile, the global grid data will be interpolated to uniform gaussian grids according to ${CLIMAT_gaussout} and saved in another set of files with suffix '_lo'.
CLIMAT_window: Similar to splitout but for any model grid configuration. Cuts out a window and treats it as high resolution area. The full grid is still treated following the value of ${CLIMAT_gaussout} (interpolated gaussian grids). Usage: CLIMAT_window="lon1 lon2 lat1 lat2", where lon1, lon2 and lat1, lat2 are the left, right and lower, upper grid indices, respectively.
CLIMAT_update_ghg: If 'CLIMAT_update_ghg=1' the values for the GHG's (CO2, N2O, CH4, CFC11, CFC12) will be updated in 'gemclim_settings.nml' according to table $gemclim/dfiles/greenhouse_gases.dat. The default is 0.
Note! This feature is only meaningful when RADIA='cccmarad' in 'gemclim_settings.nml'.




Author: Katja Winger
Last update: January 2010