CBS/ISS JOINT meeting ET-DRC and CT-MTDCF ET on Data Representation and Codes; and CT on Migration to Table Driven Code Forms, 
Montreal, Canada, 8-12 May, 2006

Photo of meeting participants thanks to Michal Weis, IBL Software

Michal Weis from IBL  presented a document to the meeting. Click here to see this paper.

Paul Heppner's HMEI report on the meeting is available below.

Comments on the meeting:

From Paul Heppner:
To watch the working of this expert group is very impressive.  There is a deliberate validation process that is employed before codes ever see the light of day, operationally.  That is not to say that mistakes aren't occasionally made.  However, when codes are proposed, they go through plenary review, where they are scrutinized and debated.  Subgroups work out more intricate detail.  The voice of the different meteorological agencies, and indeed the WMO Regions, is definitely heard through this process.   After the format of the code is expressed, it goes through a "validation" process (let's call that "testing"), then it becomes pre-operational after validation.  Final CBS approval is needed for the code to become fully operational as part of the WMO Codes – given in the Manual on Codes (WMO No. 306) and transmitted on the Global Telecommunication System (GTS).  A WMO representative, Mr. Joel Martellet, attends these meetings to ensure that the process is acceptable to the norms and practices of the WMO.

From Michal Weis :
Problems which we raised, as representative of Industry, where on confusion on BUFR - flexibility vs. standardisation. However a common conclusion has not yet been found and the problems persist. Michal Weis expressed his hope that the future will bring those answers soon, because without it, he believes real implementation of BUFR will be very difficult or even not be possible. A validation service, which was proposed in my paper, will probably be established in a near future, however it is questionable what back-end will be used, and whether it will serve for all the goals that it should.

HMEI Report
by Paul Heppner

List of Participants
Mr. Fred Branski, NOAA, USA (Chairman, CT/MTDCF)
Mr. Milan Dragosavac, ECMWF, UK (Chairman, ET/DR&C)
Mr. Joel Martellet, WMO, Switzerland (WMO Secretariat)
Dr. Weiqing Qu, BOM, Australia
Mr. Charles Sanders, BOM, Australia
Mr. Waldenio Almeida, CPTEC, Brazil
Mr. Yves Pelletier, EC, Canada
Mr. Jeffery Ator, NCEP, USA
Dr. Simon Elliott, EUMETSAT, Germany
Mr. Neil Halsey, ICAO, Canada
Dr. Olli Turpeinen, ICAO, Canada
Mr. Paul Heppner, HMEI (3SI), USA
Mr. Michal Weis, HMEI (IBL), Slovak Republic
Ms. Fang Zhao, CMA, China
Mr. Atsushi Shimazaki, JMA, Japan
Mr. Jan Willem Noteboom, KNMI, the Netherlands
Dr. Eva Cervena, CHI, Czech Republic
Ms. Sibylle Krebber, DW, Germany
Mr. William Chilambo, NMTC, United Republic of Tanzania

Discussion Items:
The meeting began with plenary (large group) discussion for Data Representation and Codes.  During these plenary sessions, various changes for GRIB2, BUFR, and CREX were presented by the experts for roundtable discussion.  These items primarily identify operational needs (e.g., the need to add new bufr table entries for atmospheric chemistry constituents, for example), or to add and correct various meteorological parameters found in existing BUFR tables.  The expert group evaluates the code table integration issues.  Matters unresolved in plenary discussion were put to subgroups for resolution, then subsequent reporting back to the plenary.

After the data codes were discussed, the next part of the meeting concentrated on migration issues and timetables for Table Driven Codes.  GRIB2 is beginning to move into operational use now.  In the United States, the National Center for Environment Prediction (NCEP) will be fully transmitting models in GRIB2 format by the end of 2006.  However, the timetable for conversion and transition at many organizations, including ICAO, is much longer.  Therefore, it is expected that GRIB1 will remain available for several years to come – probably grandfathered 3-5 years for aviation users.  However, over time, GRIB1 will eventually phase away.  Commercial vendors and meteorological organizations are encouraged to begin taking relevant GRIB2 experience now in 2006.  Datasets are available and validated.  These will mature operationally during the next couple of years.

BUFR has been operational for several years now, with some code forms (e.g., AMDAR) already expressed operationally in BUFR format on the GTS.  The schedule is for other data types to migrate to BUFR, including SYNOP and TEMP in 2006, then METAR/TAF/SPECI (aviation codes) in 2008.  These are the proposed dates when operational exchange begins.  The migration completion dates are much longer (2010 and 2015, respectively).  Effectively, this means that while operational exchange of these new BUFR forms begins within the next couple of years, the actual transmission of ascii METAR/TAF/SPECI codes for aviation on WAFS will continue for a longer period of time, perhaps until 2011.  The main point is that over the next five years, an increasing number of text-based messages that are commonly used in operational meteorology (SYNOP, METAR, TAF, etc.) will be expressed in binary (BUFR) code.  The ascii, text-based expressions of these codes will cease in the next 5 to 10 years.  This transition period should give meteorological organizations and manufacturers sufficient lead time to develop technologies to process and visualize the new data expressions for these very traditional codes! 

During this meeting, HMEI representative Mr. Michal Weis (IBL Software) presented a paper that described several issues with Table Driven Code Migration (TDCF).  Weis said that "there is a significant gap between TDCF producers and its final users at the data-processing.  Both groups are having different expectations from the software, processing, and level of knowledge of data."  Also, Weis expressed that TDCF testing is very time-consuming, owing primarily to the producer-user issues.  Many times, producers of TDCF have no possibility for cross-checking validation by other third parties.

HMEI representative Mr. Paul Heppner expressed concern to the plenary group that there are some serious gaps between what is developed in new code forms, versus end-user visualization.  Ultimately, what is expressed to the forecaster or end user is what counts on a daily, operational basis.  For example, Unknown Precipitation (UP) is a relatively new present weather type in METAR, given by automatic weather stations, yet the WMO has yet to publish a symbol for this UP present weather.  For station model plots, the user needs a symbol for UP.  He cannot be left with an impression of no symbol when UP is actually reported, because he could mistakenly think that no precipitation is occurring when indeed something is happening!  And because UP is more common around the 0C transition among solid and liquid water phases (i.e., snow, ice, freezing rain), these occurrences of UP are very critical to the aviation weather forecaster!  In METAR reports, UP is seen to be given with different Intensity and Descriptor prefixes (e.g., -UP, SHUP, etc.).  Like the other present weather types, there needs to be symbology to convey these differences to the end user.  It is not acceptable from the position of operational meteorology to report new weather types without defining appropriate, accompanying symbology.  As an outcome of this Montreal meeting, this concern about lack of UP symbology will be referred to other WMO committee for study and resolution.

Actions:
The expert group is completing a final report, which will be submitted to WMO.  Work from this group will be reviewed by the Commission of Basic Systems (CBS) in November 2006, then become operational the following November 2007. 

The WMO publishes a website to inform all interested parties about the code table migration schedules, history, etc.  The website is: http://www.wmo.int/web/www/WMOCodes.html

The community is encouraged to monitor this website for information and updates.

The experts in this group then continue on their own individual action items (test, validation, etc.) and share information with other group members in a collaborative way, so as to prepare for the next expert meeting, and also ready things for CBS.