Find in this page technical specification of the required input data and processing workflow.

Parameters

• Output time: 00 and 12 UTC

• Output frequency: between 1 and 6 hours (including step 0)

  • the model output steps should be the shortest available

• Output type:  forecast

  • only one forecast product per centre is expected for verification - the best available one for the common inter-comparison

Follow the links for exact GRIB2 WMO compliant definitions.

Encoding

All details about the required input files including sample data can be found in the page Encoding details

Sample data

• Providers are asked to provide full sample data from one model run containing all expected parameters, steps, forecast types etc.
• Naming convention for sample data:
  wave_<Organization_acronym>_<YYYYMMDDRR>_<prod/test>_<an/fc>.grib2
  acronym listed in the table below
  prod/test ... production or test version of the data
  an/fc ... analysis or forecast
• Merge all parameters from all steps for given model run and output type into one file:
  e.g. wave test forecasts from ECMWF on the 1st of September 2017, 12Z  => wave_ECMWF_2017090112_test_fc.grib2
• Put sample data on ecgate to a directory with reading permission
  • alternatively use any other downloadable location (e.g. google drive)

  • alternatively use this ftp:

    Log in to see ftp details

    
    

    host: ftp.ecmwf.int user: observations passwd: observations2013 directory: wave

    
    

• The naming convention for future production data might change

Production

• The data will be pulled via ftp from an agreed server location at the data producer's site.
• The ECMWF's acquisition system will be used for handling the production data exchange.
• Participants are committed to guarantee a steady and reliable provision of their data and to communicate any relevant changes.
• The support mailing list on provider's side will be required to allow smooth communication related to common production issues in the future.

Contact