...
Please follow instructions below to pass easily and quickly
...
development2production period of your dataset.
Progress information related to every dataset is available at TIGGE-LAM progress report pages (there will be created a detailed progress report sub-page for every model).
...
- Preparation of the first sample dataset:
- Check Archive archive specification and prepare a test dataset as close as possible to defined requirements.
- The data can be provided in GRIB1 format too. The conversion to GRIB2 TIGGE-LAM compliant files will be handled by our scripts.
- If possible please keep the structure of your input files names and content as <PREFIX>_<NN>+<HHHH>.grb where
- PREFIX is arbitrary
- NN = 00 for control forecast (if exists)
- NN = 01, 02..maxN where maxN is the last perturbed member of the ensemble; NN format should be of type %02d
- HHHH = 0003, 0006..maxH where maxH is the last forecast lead time. HHHH format should be of type %04d
- all defined parameters available for the given type of forecast are in one file
- e.g. LAEF_4tigge_00+0030.grb is the control forecast at step +30h
- Include ensemble meta-data information in the input GRIB files (if below GRIB keys couldn't be added to the input GRIB files, e.g. to their "Local section" parts, it must be possible to derive that information from the file names):
- size of ensemble (GRIB key numberOfForecastsInEnsemble)
- ensemble member's number (GRIB key number)
- type of forecast (GRIB key type = pf/cf i.e. perturbed/control)
- The static fields (orography and land-sea mask) should be provided only at step 0 in control forecast or in a separate file if control forecast doesn't exist.
- If you have an account at ECMWF put the test data anywhere in provided IT facilities and make them readable for everybody or at least for emos user.
- If you don't have an account at ECMWF use your own FTP server and provide the connection details (alternatively some FTP server can be created at ECMWF for data pushed by providers but to use the providers' FTP account is a preferred way as well for following operational data exchange)
...
- The input data converted to TIGGE compliant grib2 files will be stored in TIGGE-LAM archive in the test mode.
- The daily monitoring and checking of models' data availability is possible using tools linked here in the Resources section.
- The end of test phase is achieved by successful UAT (user acceptance testing) done by data provider:
- The data can be retrieved using provided user interfaces.
- The data values and pattern are correct (some basic monitoring tools provided are listed above).
...