At the moment the ECMWF's interpolation package cannot handle rotated latlon projection in which 4 TIGGE-LAM datasets are coded :
It means that at the moment users cannot ask for other than default grid for those 4 datasets (situation will change by 2016 after new interpolation package MIR is finished and implemented). Till then users are advised to use any other external interpolation package which can handle the rotated latlon. One of them can be COSMO package versus described below (thanks to Unknown User (dcesari@arpa.emr.it))
COSMO open source verification/interpolation package versus
The package including installation and users' instructions can be downloaded at
ftp://ftp.smr.arpa.emr.it/incoming/dav/versus
In order to install the binary package, which will work on any reasonably recent linux x86_64 platform, you need to do the following operations:
# go to the directory where you wish to install the software
tar -zxvf path/to/versus_extra_software-1.4_unibin.tar.gz
cd ves
./install_from_bin.sh
at this point the installation is complete and you will have, in your home directory, a file called "ves_profile". When you need to use the package in a terminal session or in a batch script you have to "source" once that file, assuming you use bash/ksh, in the following way:
. ~/ves_profile
after this operation you have access, within that session, to all the programs of the package. For interpolating grib data to a regular lat/lon grid you can use the following command:
vg6d_transform --trans-type=inter --sub-type=bilin \
--type=regular_ll --nx=... --ny=... \
--x-min=... --y-min=... --x-max=... --y-max=... \
inputfile outputfile
where, of course you have to fill the parameters of the target grid, just start vg6d_transform without arguments to see the instructions. The package contains many other tools and utilities for managing grib and observed data.
Unknown User (dcesari@arpa.emr.it) (2015-05-06) Metoffice Mogreps grib messages are coded with an ending longitude >360, it is clear that this has been done with the good intention to make endLongitude>startLongitude when the grid crosses the zero meridian, but this seems to violate the MO rule stating that 0<=lon<=360 and I am afraid that it will generate an unexpected situation in versus software so that it can't be handled at the moment. latitudeOfFirstGridPointInDegrees = -3.77995; longitudeOfFirstGridPointInDegrees = 354.911; latitudeOfLastGridPointInDegrees = 7.16005; longitudeOfLastGridPointInDegrees = 363.311; It will be tried to make sure that in next version of libsim/versus extra software all the reasonably odd cases for coordinates are taken care of, including this one. |