Documentation of the Changes in the COSMO-Model
Version 5.06a

Contents:

1. Skin Temperature Formulation
2. Latest Modifications to the Turbulence Scheme
3. Technical Changes and Bug Fixes
4. Changes to the Namelists
5. Changes of Results

1. Modifications to Physical Parameterizations

(by Jan-Peter Schulz)

A skin temperature formulation (Viterbo and Beljaars 1995, Schulz and Vogel 2017) was implemented in TERRA in order to represent the vegetation in the surface energy balance. It provides an additional energy budget including a skin (or canopy) temperature. Furthermore, the insulating effects by the vegetation at the sub-canopy level are represented, as well. This scheme allows to enhance the amplitude of the diurnal cycle of the simulated surface temperature in TERRA. At the same time, the amplitudes of the diurnal cycles of the simulated soil temperatures are reduced.

This parameterisation provides the skin temperature SKT as new (output) variable.

The scheme is switched on with the /PHYCTL/ namelist switch itype_canopy=2.

It requires one additional external parameter, the skin conductivity SKC, which is provided by EXTPAR through INT2LM (from Version 2.06 on).

The use of SKC is controlled by the new /PHYCTL/ namelist switch cskinc. For a negative value, cskinc = -1.0, the external parameter field SKC is read in and used. For case studies or other testing purposes, a globally constant value can be used, instead. Reasonable values are cskinc ∈ [10.0, 1000.0], the default value is cskinc = 30.0.

Additionally, the scheme requires the numerical stability parameter cimpl, specified in /PHYCTL/ as well. The default value is cimpl = 120.0, this value works well for COSMO domains with 2.2 km mesh size. For coarser grids and longer timesteps, lower values are appropriate, e.g. cimpl = 30.0 as in ICON. Reasonable values are in principle cimpl ∈ [4.0, 200.0].

Back to Contents

2. Latest Modifications to the Turbulence Scheme

(by U. Schättler)

Due to the unification with ICON there were some more modifications to the turbulence scheme.

• turb_diffusion.f90:
  • Eliminated global field eprs (is now a argument to SR turbdiff).
  • Eliminated some fields / real variables, which are not used any more.
  • Only compute my_thrd_id if OpenMP is used.
  • Bug fix for the Intel FORALL statement, which still used istartpar, iendpar instead of ivstart, ivend.
• turb_interface.f90:
  • Need the fields z3drhon, z3dvari with attribute SAVE, otherwise the definition/allocation status gets lost.
  • Do not recompute rhon for vertical diffusion of scalar (mass grid point) variables (for that set lzturatm=.TRUE.), but have to compute it for wind grid points (for that set lzturatm=.FALSE.).
• turb_transfer.f90:
  • Eliminated some fields / real variables, which are not used any more.
  • Only compute my_thrd_id if OpenMP is used.
• turb_utilities.f90:
  • Info print for accelerators only for COSMO (ICON needs a different way for info prints).
  • Need some fields as INTENT(INOUT), not (OUT) for ICON.
• turb_vertdiff.f90:
  • Remove initializing rair_tar to 0.0 (is not yet needed).
  • Still need dp0 in the argument list for recomputing rhon for wind grid points.
  • Eliminated some fields / real variables, which are not used any more.
  • Only compute my_thrd_id if OpenMP is used.

Back to Contents

3. Technical Changes and Bug Fixes

• organize_physics.f90:
  Need the field zttens_conv with attribute SAVE, otherwise the definition/allocation status gets lost (or put it to data_fields as ttens_conv_gscp: temp tendency from microphysics).
• src_obs_gnss.f90:
  Have to initialize Ncols with 0 at the beginning of SR organize_gnss_std.
• src_obs_rad.f90:
  Initialize variable yroutine in SR prepare_rttov_input.
• src_sat_rttov.f90:
  Initialize variable yroutine in SR prepare_rttovlevels.
• src_setup.f90:
  • Allocation of isubpos on accelerator can only be done on compute PEs.
  • Set nproma_d for accelerators to -1.
• src_sfcana.f90:
  Use variable ydatname as argument to SR close_file and eliminated ypath.

Back to Contents

4. Changes to the Namelists

Back to Contents

5. Changes of Results

None.

Back to Contents