[ferret_users] missing value and scale factor

[Bruno Levier <bruno.levier@xxxxxxxxxxxxxxxxx>]

Hi Ferreters

I want to plot a velocity field, but some values appear as "missing" and 
they shoud not. Here's an example :

     FERRET v6.2
yes? use pgs.nc
yes? list u
               19.416W 19.388W 19.361W
                  1       2       3
 27.388N / 3:  0.0262  0.0262  0.0262
 27.361N / 2: -0.0043    .... -0.0183
 27.333N / 1: -0.0647 -0.1044 -0.1154

The value u[2,2] is missing.

The "ncdump" of the file gives this :

    short u(time_counter, latitude, longitude) ;
        u:long_name = "Eastward velocity" ;
        u:standard_name = "sea_water_x_velocity" ;
        u:units = "m s-1" ;
        u:unit_long = "Meters per second" ;
        u:_FillValue = -32767s ;
        u:add_offset = 0. ;
        u:scale_factor = 0.000610370188951492 ;
        u:valid_min = -4984s ;
        u:valid_max = 4528s ;
        u:cell_methods = "area: mean" ;

 u =
  -106, -171, -189,
  -7, -20, -30,
  43, 43, 43 ;

So, the short value -20 ( corresponding to u[2,2] ) should give 
-20*0.00061037 = -0.01221 m/s (multiply the short value by the scale 
factor), but Ferret replaces this value by a missing one.
In fact -20 is equal to the product -32767*0.00061037. It seems that 
Ferret multiplies the FillValue by the scale factor, so that values 
equal to -20 are flagged as bad.

This is unclear for me. Any suggestions ?

Thanks in advance,

Bruno Levier