Re: how to fill the data in the missing value

[Ansley Manke <Ansley.B.Manke@xxxxxxxx>]

Hi,
I think there is information to get you started with this under
"filler (missing value)" in the Users Guide.  There is an averaging
fillers which let you choose the size of the averaging window;
a linear interpolation filler, and a nearest-value filler.

I am not sure how you might switch from time to spatial interpolation;
perhaps after you work with the filling transformations, if you haven't
found a way to do what you want, you might ask the users list again.

Ansley Manke

Namba Takaya wrote:

> Dear Sir
>
> I have a netcdf data like bellow.
>
> bash-2.05b$ ncdump -h upstr1.nc
> netcdf upstr1 {
> dimensions:
>         LON91_110 = 20 ;
>         LAT91_110 = 20 ;
>         TT = 12 ;
> variables:
>         double LON91_110(LON91_110) ;
>                 LON91_110:units = "degrees_east" ;
>                 LON91_110:modulo = " " ;
>                 LON91_110:point_spacing = "even" ;
>         double LAT91_110(LAT91_110) ;
>                 LAT91_110:units = "degrees_north" ;
>                 LAT91_110:point_spacing = "even" ;
>         double TT(TT) ;
>                 TT:units = "DAY since 0000-01-01 00:00:00" ;
>                 TT:time_origin = "01-JAN-0000" ;
>         float UPSTRR(TT, LAT91_110, LON91_110) ;
>                 UPSTRR:missing_value = -1.e+034f ;
>                 UPSTRR:_FillValue = -1.e+034f ;
>                 UPSTRR:long_name = "UPSTR37_CLIM[G=XXYYTT]" ;
>                 UPSTRR:history = "From upstr.mean.nc" ;
>
> (in ferret)
> yes? use upstr1.nc
> yes? show data
>      currently SET data sets:
>     1> .\upstr1.nc  (default)
>  name     title                             I         J         K         L
>  UPSTRR   UPSTR37_CLIM[G=XXYYTT]           1:20      1:20      ...
> 1:12
>
> There are some missing values of UPSTRR.
> I would like to fill(interpolate) the values in stead of missing values and
> I would like to make none of missing values in UPSTRR.
>
> Of course, it depends on how to interpolate.
> For exmaple, it is possible that I do "at first time(T) interpolation then
> next X and Y spacial interpolation" or "X and Y spacial interpolation then
> time interpolation".
>
> Then is it possible to fill the values in the all missing valuses to sweep
> out the missing value?
>
> Could somebody tell me how?
>
> Best Regards
>
> Takaya Namba
>
> _________________________________________________________________
> $B%O%$%;%s%9$J>&IJ$r5$7Z$K9XF~(B MSN $B%7%g%C%T%s%0(B   http://shopping.msn.co.jp/

--
Ansley Manke  Pacific Marine Environmental Laboratory  Seattle WA  (206)526-6246