RPN standard files

For the end-user, the appearance of the FST is that of a flat-file database structure. A single file contains numerous records which hold fields on 1, 2 or 3D data grids and self-describing information about the fields (header).

A thorough set of FST-related routines are contained in the RPN Library (armnlib). These routines provide easy access to the file contents. They can be called from FORTRAN or C routines.

A complete description of the FST file format and the utilities which can be used to manipulate and view their contents is available online at www.cmc.ec.gc.ca/rpn/utilities.

Sometimes the files are archived in an cmc-archive file. If this is the case they need to be unarchived first.

How to look at FST files


voir is the most simple way to see the content of an FST file (also xvoir)
It prints the headers for all of the records of the given RPN standard file:

  voir -iment fst-file [-style]

The -style option decodes the level and date information.


r.fstliste is available if you need a bit more information than what is provided by voir.
It can be used in batch mode to extract specific or all of the record parameters of a RPN standard file record(s) based on a criteria given through the selection keys:

  r.fstliste -izfst fst-file [-datev -vdatev -ip1 -typvar -nomvar -col ...]


xrec is a X-window visualization program used to display 2D (horizontal or vertical) meteorological fields stored in FST files. Usage:

  xrec

or

  xrec -imflds list_of_FST_files


How to modify FST files

The two main tools to modify FST files are called editfst and r.diag. And of course, you can also write your own C or Fortran programs.


editfst (also called editfst2000) is a utility used for editing and copying records from RPN standard files into a new or existing RPN standard file. It can do a straight (complete) copy of the input file or it can copy records selectively as indicated from the standard input or from a file of directives named in the "-i" key. One can select (via desire), exclude (via exclure) or zap records by their:
  • TYPVAR: type of variable; 'A'(analysis),'P'(forecast),'X'
  • NOMVAR: 4 caracter variable name
  • ETIKET: 12 caracter label recorded from the original model run
  • DATE: CMC date timestamp
  • IP1: the level of the field (in pressure,eta,meters,other)
  • IP2: the hour of the forecast (rounded off if not exact)
  • IP3: value of 0 unless otherwise modified
    • Example:
    You could use it to select i.e. temperature (TT) and winds (UU,VV) at 500 hPa:

      echo "desire (-1,['UU','VV'], -1 ,-1 ,[500.,MBAR])" | editfst -s input_file -d output_file

    Note:
    editfst appends to its output file. This means that if editfst is used more than once writing into the same output file, the output of the second set of commands will be appended to the output of the first set.


    r.diag (also called r.diag2000) is a toolkit used to manipulate FST or CCCma files. As editfst, it can be used to select or exclude specific records but in addition to this it also contains a large set of functions which can be used to manipulate the files.

    'r.diag' is always used in the way:

      r.diag function input_file(s) [ output_file(s) key(s) ]

    To get information about a specific function simply type:

      r.diag function


    Example:
    Selecting temperature (TT) and winds (UU,VV) at 500 hPa as above:

      r.diag select input_file output_file_tt output_file_uu  output_file_vv  -name TT UU VV -lv1 500

    Notes:


    Writing your own C and Fortran programs
    There is a number of functions which can be called in a C or Fortran program to read and write FST files, and they are documented in the RPN Standard File Functions. These functions are part of the library 'librmn'.

    There are functions to:
    Have a look here on how to compile.




    Author: Katja Winger
    Last update: June 2008