09.11.2021
- Adaptations to latest COSMO-ART version
- Fixing the Block Copies
- Technical Changes and Bug Fixes
- Changes to the Namelists
- Changes of Results
(by Ulrich Schättler)
Due to the introduction of the tiles in the COSMO-Model, the ART code had to
be adapted. This is done in version ART3.2.
There have also been modifications to the module
mo_art_rad.f90, which prepares the computation of blocked fields for radiation feedback
in COSMO-ART. Calculation of fields in ijk format is now done in an extra step
before entering the block-loop in organize_physics.f90. Additional fields are added
to the register-copy calls in radiation_interface.f90.
Back to Contents
(by Ulrich Schättler)
Splitting the block loop for running the physical parameterizations in
organize_physics.f90 to implement a coupling for the CLM changes the results
of the forecast, which points to problems when doing the "copy-To/From-block"
for the blocked variables.
The problem is in the radiation_interface.f90 for the variables that are also
used in TERRA and in SNOWPOLINO.
Most of the output variables from the radiation are only used in the diagnostics
and for output, so the ij-variables are set directly. Only a few variables are
used in TERRA and / or in SNOWPOLINO. For these variables we also set the blocked
variables directly in addition to the ij-variables. To indicate the necessity of
updating these variables we also include them in the "copyFromBlock" mechanism.
Here we corrected two bugs now:
- TERRA uses the variables thbs (thermal radiation) and sobs, pabs (solar radiation).
By accident, also the variable thbs was added to the "radso" list for solar
radiation, but as it is not set when only running the solar radiation, wrong
values were given back to the ij-variables.
The forecast results are only affected when splitting the block loop in
organize_physics, because TERRA then uses the wrong results. When running the
block loop without splitting (as it is done up to now), only the output variables
for THBS_RAD are affected.
- SNOWPOLINO uses the variables lwd_s, lwu_s (thermal) and
swdir_s, swdifd_s, swdifu_s (solar):
Up to now the blocked variables of these fields were not set in radiation_interface,
so up to now SNOWPOLINO uses the value of these fields from the last time step.
Splitting the block loop and copying the ij-fields to the blocked fields has
the effect, that now the correct value computed in the radiation is given
to SNOWPOLINO.
This is fixed now by also setting the blocked variables of these fields in
radiation_interface.f90.
Back to Contents
- environment.f90:
Creation of END file in model_abort added for MESSy.
- organize_data.f90:
Require at least one GRIBOUT block, so minimum value for ngribout is 1
(before, only ngribout ≥ 0 was checked).
- radiation_interface.f90:
Set blocked variables for some radiation fields, which are used in snowpolino.
- src_output.f90:
Added additional tile-dimension to zvarlev in EULAG code.
- src_setup.f90:
Increased allowable interval for fac_rootdp2 from 0.5 to 5.0 (was 1.5 before).
- src_setup_vartab.f90:
Added additional tile-dimension for EULAG fields.
Back to Contents
| Group |
Name |
Meaning |
Default |
| /IOCTL/ |
ngribout |
CHANGED |
Minimum value for ngribout now is 1. |
1 |
| /TUNING/ |
fac_rootdp2 |
CHANGED |
Increased allowable interval for fac_rootdp2
from 0.5 to 5.0 (was 1.5 before). |
1 |
Back to Contents
NONE
Back to Contents