Notes on the process of using nadcon (using the Alaska data and
generating a 15 minute grid as the example)

1.  The process starts with acquiring or editing the alaska.dat file.

2.  The allwnd program is run to window the data and change its
    format.  Allwnd is run interactively.  The correct answers
    for Alaska are:

    (1) 2                => station data
    (2) alaska.dat       => input file name
    (3) alaska.wnd       => output (windowed) file name
    (4) 50,74            => minimum and maximum latitudes in degrees
    (5) 128,191          => minimum and maximum west longitudes in degs

    The correct answers for Conus are:

    (1) 2                => station data
    (2) conus.dat        => input file name
    (3) conus.wnd        => output (windowed) file name
    (4) 24,50            => minimum and maximum latitudes in degrees
    (5) 66,126           => minimum and maximum west longitudes in degs

    The screen output consists of

    1. The first three records in the input file
    2. The first four records in the window not including the
       first three (two?) records in the file
    3. The windowing information including maximum and minimum shifts.
       This information must be saved, say by redirecting the
       output.  This information can be used to edit the alaska.dat
       file if it is not clean.

    The output of the allwnd program may be used to edit the .dat
    file (e.g. alaska.dat).

3.  When the .wnd file exists, the convdam program is run.
    It creates the .bin file.

    The first step here is to split the .wnd file into segments of
    6000 or fewer records using the Unix split command.  e.g.

       split -6000 alaska.wnd

    The convdam.in file must then be edited.  The input records
    consist of:

    1) The number of xaa, xab, xac, ... files generated by the
       split command
    2) the names of the x?? files, one per line
    3) the output binary file name. e.g. alaska.bin
    4) The output file header record.  This should contain the
       date of the run.
    5) The number of variables on each record.  This must be the
       number 9.
    6) The FORTRAN format for reading the data.  This record must be:
       (A38, 4(2X, F14.9), 4(2X, F8.3) )
    7) The output group headers, one record for each input x?? file.
       This is usually the number of the input file. i.e. "1" for the
       xaa file, "2" for the xab file, etc.
    8) n (for no) - this record must consist of this character.
    9) n (for no) - this record must consist of this character.

    If there are 6000 or fewer records in the .wnd file then the
    split command does not need to be used.  Instead, input records are

    1) 1
    2) alaska.wnd
    3) alaska.bin
    4) header info - same as above
    5) 9
    6) (A38, 4(2X, F14.9), 4(2X, F8.3) )
    7) 1
    8) n
    9) n

    The convdam program should only take several minutes to execute.
    It is interactive.  Thus the Unix command is:

    convdam < convdam.in

    The output generally does not need to be saved.  The x?? may
    be removed.  Keep the *.wnd file as it is used later.

4.  Next, the two parameter files and the two input files for the
    mincurv program need to be created/edited.
    (The new proceedure needs to create only the seconds of arc shift
     files (.las and .los) and not to create the meter shift files
     (.lam and .lom).  However, they can be made for plotting.)

    The two parameter files are named: maplas.par, maplos.par.
    These files have the following information on the records
    (starting with column 44):

    1) The first line has the name of the binary input program.  This
       is the output file from the convdam program.  e.g. alaska.bin
    2) The name of the gridded output file. e.g. alaska.las for the
       maplas.par input file.
    3) The header line for the output
    4) The column in the input which contains the X data.
    5) The column in the input which contains the Y data.
    6) The column in the input which contains the Z data.
       For latitude shifts in meters (.lam) it must be a "6" (not used)
       For longitude shifts in meters (.lom) it must be a "7" (not used)
       For latitude shifts in seconds (.las) it must be a "8"
       For longitude shifts in seconds (.los) it must be a "9"
    7) The longitude in degrees of the origin (i.e. the lower left
       hand corner).
    8) The latitude in degrees of the origin (i.e. the lower left
       hand corner).
    9) The increment of the grid in decimal degrees. i.e. for a 15
       minute grid, the increment is 0.25 degrees.
   10) The angle which the grid should be rotated from the X axis.
       This must be a zero.
   11) The number of columns in the grid.  This can be calculated from
       the left (west) and right (east) longitudes of the grid and the
       grid increment all in degrees. i.e. (left - right)/increment+1
       Alaska => 63/inc+1, Conus => 60/inc+1
   12) The number of rows in the grid.  This can be calculated from
       the upper (north) and lower (south) latitudes of the grid and the
       grid increment all in degrees. i.e. (upper - lower)/increment+1
       Alaska => 24/inc+1, Conus => 26/inc+1

    An example for the maplas.par file is:

       1. FIELD DATA FILE: ......................alaska.bin
       2. GRID FILE: ............................alaska.las
       3. HEADER NOTE: ..........................coord shifts NAD27
       4. VARIABLE NUMBER OF X: .................              2
       5. VARIABLE NUMBER OF Y: .................              3
       6. VARIABLE NUMBER OF Z: .................              6
       7. LOWER LEFT GRID CORNER COORD: XO= .....     -191.00000
       8.                               YO= .....       50.00000
       9. X AND Y INCREMENT: ....................         .25000
      10. ROTATION ANGLE FROM X-AXIS (DEG): .....            .00
      11. NUMBER OF GRID COLUMNS: ...............            253
      12. NUMBER OF GRID ROWS: ..................             97
      (blank line)

    Note that the only difference between the parameter files will
    be the sixth record.

    The input files for the mincurv program are named:
    input.las.min, and input.los.min.  They each have
    three records.  The only difference between the input files will
    be the second record.

    1) 1 -- it indicate that there is an input parameter file.
    2) the name of the input parameter file.  e.g. maplas.par
       for input.las.min input file.
    3) n  -- to indicate that the program should quit here.

    We now nave the two parameter file, the two input files,
    and the input binary data file (e.g. alaska.bin, the output of
    the convdam program).

    The mincurv program is executed two times, once for each
    set of input and parameter files.  e.g.

    mincurv < input.las.min > output.las.min
    mincurv < input.los.min > output.los.min

5.  The new proceedure is to run naddif on the allwnd output (i.e.
    the conus.wnd, alaska.wnd, etc file).  naddif has been modified
    to read this input format.  the file type in 3.

5a. The output of the naddif run can be analyzed with the dstat program.

6.  To iterate the size of the grid go back to step 4.