NOTE: There has been a DWD-internal intermediate version 2.04a (small bug fix in computing pp_lm) whose change log has been integrated below!
Three additional fields can be provided in the external parameter data set for COSMO describing the Kinne aerosol climatology (2013). These fields provide monthly mean values of total aerosol optical thickness, single scattering albedo and asymetry parameter for the 8 wavelengths of the COSMO radiation scheme as well as for the 550 nm wavelength. However, these parameters are only available in the NetCDF format of the external parameters.
The short names in COSMO external parameter files are: AOT12, SSA12, ASY12
The resulting time-interpolated fields in the laf-files have the shortNames TAU_KINNE, SSA_KINNE, and G_KINNE.
In INT2LM this can be activated by the new option itype_aerosol = 3 (/CONTRL/).
NOTE:
The implementation of these fields in the COSMO-Model
itself and its radiation scheme is not yet available in the official version!
It is implemented in a test version of the T2(RC)2 project!
ECMWF provides operational aerosol forecasts from a relatively coarse (~60 km resolution) deterministic IFS instance. The data are provided as (dry) mass mixing ratios for 11 categories. In IFS, the optical properties of the soluble species (such as sea salt), depend also on relative humidity, which enables to represent the direct radiative effects of haze in the COSMO-model.
This can be activated by the new option itype_aerosol = 4 (/CONTRL/).
NOTE:
The implementation of these fields in the COSMO-Model
itself and its radiation scheme is not yet available in the official version!
It is implemented in a test version of the T2(RC)2 project!
In the scope of the PerduS project, daily mineral dust forecasts with ICON-ART are produced. ICON-ART provides mass-mixing ratios of three log-normally distributed modes.
This can be activated by the new option itype_aerosol = 5 (/CONTRL/).
NOTE:
The implementation of these fields in the COSMO-Model
itself and its radiation scheme is not yet available in the official version!
INT2LM can now be used with ecCodes as well as with grib-api.
For that a problem in src_lm_output.f90 was fixed: Modification in SR makeapi2_pdt for eccodes to get a proper interpretation of indicatorOfUnitOfTimeRange and stepUnits:
Due to a problem in former versions of grib_api, the productDefinitionTemplateNumber for generalVertical level type was set after all keys for the vertical level. With eccodes this led to a mis-interpretation of the indicatorOfUnitOfTimeRange, if this was not set to hours, but to minutes. This mis-interpretation could be solved by also setting the computed key stepUnits, but also by setting the productDefinitionTemplateNumber for all products before setting the keys for vertical level and time range coding.
NOTE:Results are not changed.
Modifications were performed to also check the uuidOfVGrid, if HHL fields are read from an extra icon_hhl-file. In this new version the namelist switch lcheck_uuidOfHGrid should (must) be set to .TRUE..
NOTE:Results are not changed.
A new namelist switch lextpar_lm_only has been introduced in group /CONTRL/. If set to .TRUE., only the external parameters of COSMO are processed for the initial time (hstart) and written to the laf-file. There is no need to provide the atmospheric fields from the coarse model for the same time, i.e., only the external parameter files from COSMO and the coarse model are needed at initial time.
This enables to get the time-dependent external parameters (climatologies, PLCOV, LAI, ...) for a specific time independent of any atmopheric coarse model data. It is possible, however, to process atmospheric fields for boundary files (lbff) as usual, if lboundaries=.TRUE.
NOTE:Results are not changed.
NOTE:Results are not changed, apart from avoiding crashes.
In this case the boundary data for PLCOV, LAI, ROOTDP and other time-dependend fields were all 0.0, because either no external parameters were written (PLCOV, LAI) or not saved for the whole model run (ROOTDP).
This has been corrected in external_data.f90 and src_read_ext.f90.
NOTE:Otherwise, results are not changed.
The field operation
p_lm(:,:,:) = pp_lm(:,:,:) + p0_lm(:,:,:)
needs an explicit vertical dimension range, because this can be different for p_lm(:,:,1:kedim) respectively pp_lm(1:,:,1:kedim) and p0_lm(:,:,1:kelm)!
Now it is
p_lm(:,:,1:kelm) = pp_lm(:,:,1:kelm) + p0_lm(:,:,1:kelm)
Explanation:
If the input model has less vertical levels than COSMO, kelm is equal to kedim
and no problems arise. Otherwise, kedim > kelm applies and p_lm can receive
crappy values from p0_lm for levels > kelm, or INT2LM may crash with a
floating point exception depending on compiler, number of processors and number of gridpoints.
There is one new namelist variable and additional options for an existing variable:
| Group | Name | Meaning | Default | |
|---|---|---|---|---|
| /CONTRL/ | lextpar_lm_only | NEW | If .TRUE. for the initial time only COSMO external parameters are processed. However, it is possible to process atmospheric fields for the boundary files as usual if lboundaries=.TRUE. | .FALSE. |
| itype_aerosol | additional options |
| 1 | |
There are no changes of results.