Hei,
in grib, clouds are described as low, medium and high clouds, e.g.
73,74,75. Those are described by phenomenon, e.g.
high cloud type: Clouds of genera Cirrus, Cirrocumulus and Cirrostratus.
low cloud type: Clouds of genera Stratocumulus, Stratus, Cumulus, etc.
medium cloud type: Clouds of the genera Altocumulus, Altostratus, etc.
(see
In CF, this can currently only be expressed by
cloud_area_fraction_in_atmosphere_layer and a not very well defined
'vertical' parameter, e.g. by sigma:
http://www.ecmwf.int/products/data/archive/data_faq.html#clouddefinitions
When translating from grib to CF, this is not very satisfying: Either I
add some dummy sigma-values, which will look like I have a model with
just three levels, or I use 3 variables, all with the same
'standard_name="cloud_area_fraction"' which looses some useful
information. In both cases, the data-user will need to know something
which cannot be expressed by CF.
Looking at the page
http://cf-pcmdi.llnl.gov/documents/cf-standard-names/ecmwf-grib-mapping
there are mentioned 3 'standard_names' which are not in the
standard_name-table yet, and I propose to do so:
low_cloud_area_fraction
medium_cloud_area_fraction
high_cloud_area_fraction
in additon (though this is not in grib), I would like to add
fog_area_fraction (or surface_cloud_area_fraction).
Best regards,
Heiko
Received on Wed Apr 25 2012 - 03:49:06 BST
