Documentation of the Changes in INT2LM
Version 2.06

Contents:

1. Slope of Orography
2. Skin Conductivity
3. Refactoring lockfile Mechanism
4. Interpolation of hhl_in to hhl_gl
5. Treatment of Interpolation / Computation of a Control Geopotential
6. Changes to Output using NetCDF
7. Technical Changes and Bug Fixes
8. Changes in the Namelists
9. Changes of the Results

1. Slope of Orography

(by Guy de Morsier)

For the modifications regarding the ground water runoff in the soil model TERRA (by Linda Schlemmer), an additional external parameter field, the slope of orography (S_ORO) is needed. This field is provided by EXTPAR and given to the COSMO-Model laf-file through INT2LM.

To activate reading S_ORO and writing it to the laf-file, a new namelist switch l_s_oro (Default: FALSE) has been added in /CONTRL/.

Back to Contents

2. Skin Conductivity

(by Jan-Peter Schulz)

For the skin temperature formulation added in COSMO-Model 5.06a, an additional external parameter field, the skin conductivity (SKC) is needed. This field is provided by EXTPAR and given to the COSMO-Model laf-file through INT2LM.

To activate reading SKC and writing it to the laf-file, a new namelist switch lskinc (Default: FALSE) has been added in /CONTRL/.

Back to Contents

3. Refactoring lockfile Mechanism

(by Guy de Morsier)

Using the lockfile mechanism is an alternative approach (compared to READY files) to wait for input files, which is used by MeteoSwiss. It has been implemented into INT2LM some time ago but is now more similar to the mechanism used in the COSMO-Model.

Files are not written directly to a file called filename, but first to a file with name filename.lck. Only if the file is fully written and is closed, the name is changed to filename. For this the (non-standard) Fortran intrinsic RENAME is used. But this intrinsic nowadays is offered by (nearly) all compilers.

Using lockfiles is activated by setting llockfiles = .TRUE. in namelist /DATA/

Back to Contents

4. Interpolation of hhl_in to hhl_gl

(by Ulrich Schättler)

Computation of reference atmosphere and hhl_in / hhl_gl for GRIB2

Instead of re-computing hhl_gl for the fine orography grid, just interpolate it from hhl_in (because in GRIB2 and for SLEVE there are not all necessary meta data available to re-compute).

• Up to now:
  • When using GRIB2, not all meta data are available to compute the height of half levels for the incoming COSMO grid (hhl_in), therefore we read the full hhl_in field.
  • Then we try to determine the vcflat of that field to compute the height of half levels for the fine COSMO grid and the incoming vertical grid (hhl_gl). This is only possible for ivctype=2.
  • But this is not enough, if the incoming COSMO grid has the SLEVE coordinate. More meta data are missing then.
• New method:
  • Interpolate all levels from hhl_in to hhl_gl. This is also done for ivctype = 2 now, because we can avoid recomputing the vcflat then.

NOTE:
This changes the results of the atmospheric variables slightly, but only up to numerical noise. And only if incoming COSMO data in GRIB2 are used.

Back to Contents

5. Treatment of Interpolation / Computation of a Control Geopotential

(by Guy de Morsier)

For the computation of the surface pressure and then computing full pressure and splitting the relative humidity, a control geopotential level from the driving model is needed. If this is not available, it is computed by other input fields.

But this can lead to noisy pressure fields, therefore a switch has been implemented, so that the user can choose, whether such a computed control geopotential is used or not.

New namelist variable in /GRID_IN/: lcontrol_fi: Default: .FALSE.

• if .TRUE.: check, if the control geopotential level fic_in has been read, if it has not been read: abort
• if .FALSE: behaviour up to now: if the control geopotential level fic_in has not been read, it is computed by other input values.

Back to Contents

6. Changes to Output using NetCDF

(by Guy de Morsier)

NETCDF4 has been introduced as optional output format. It can be activated by ylm_form_write = 'nc-4'.

Possibility to compress NETCDF4 files:

• A new component compress has been added to the structure ar_des_lm (in data_int2lm_io.f90). With this component it is possible to have a compression for NETCDF4 files.
• It can be set for all I/O fields, but at the moment it is .FALSE. for all fields.

Changes for NETCDF coding of HORIZON.

Back to Contents

7. Technical Changes and Bug Fixes

• io_utilities.f90:
  
  • Introduction of NETCDF4 ('nc-4') for ylm_form_write and wait mechanism.
  • Modification for treatment of HORIZON and scaling of ncdf files.
• src_memory.f90, src_cleanup.f90:
  Removed all conditional memory allocation / deallocation with IF (linitial...). This caused several problems when setting lbdclim .AND. .NOT. linitial (for COSMO hindcasts).
• src_coarse_interpol.f90:
  Set negative incoming values for w_i_in to 0.0 (occured from IFS).
• src_decomposition.f90:
  
  • Subroutine check_iconsub_grid:
    Moved setting of logical print_hint into an IF-statement, where it belongs. Otherwise a warning is issued very often.
  • Subroutine decompose_cm:
    Bugfix for decomposition in case of lcm2lm: automatic adaption of nzbounds_e, nzbounds_w, nzbounds_n, nzbounds_s (BR)
• src_gribtabs.f90: Added continuation to lines with more than 132 characters (DR).
• src_lm_output.f90:
  
  • Added specification of productDefinitionTemplateNumber for ICON-ART mineral dust (DR).
  • Changes for pollen ALN,AMB,BET,POA coding (GdM).
  • Transfer levtyp for T_S from 111 (depthBelowLand, GRIB1) to surface (GRIB2) (US).
• src_namelists.f90:
  Automatic adaptation of itype_profiles_vert_interp in case of lcm2lm: For lcm2lm the variable itype_profiles_vert_interp is set to 1 (if it would be 2).

Back to Contents

8. Changes of the Namelists

NOTE:
There was another change in the namelist group /CONTRL/:
For lcm2lm = .TRUE. (which is yinput_model = 'CM') and itype_profiles_vert_interp = 2, the type of vertical interpolation is reset to 1 again: itype_profiles_vert_interp = 1.

Back to Contents

9. Changes of the Results

Interpolating hhl_in fields to hhl_gl instead of computing hhl_gl with (perhaps) available meta data changes the results numerically. But this is only when using GRIB2.

Back to Contents