Re: intepolating scattered cdf file

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

Hi Hemerson,
The scat2grid functions are intended to put data that is not
already on any axis or grid, onto a grid that you define. The
first two arguments to the scat2grid functions need to be
1-dimensional lists the same length; they describe the locations
of the points to be interpolated onto the grid.  These functions
expect a list of function values at a set of x, y locations ( x, y,
f(x,y) )

If you have data that's already on a grid, you can define new axes
and regrid to those new axes with ordinary Ferret regridding
operations.

Here are a couple of examples. First, use some data that has a
regular Y axis and an irregular X axis, regrid to a regular X axis
and a new Y axis, but keep the Z axis.  On the SHOW DATA
commands, you can see which axes are irregular.  Under the # pts
column note the the letter i for irregular or r for regular.

yes? USE myfile
yes? SHO data
     currently SET data sets:
    1> ./myfile.nc  (default)
 name     title             I      J      K      L
 TEMP     TEMPERATURE      1:10   1:40   1:20   ...

yes? SHO GRID temp
    GRID GLR1
 name       axis      # pts   start        end
 XAXL       X         10 i   123          222
 YAXL LATITUDE        40 r   19.5S        19.5N
 ZAXL DEPTH (m)       20 i-  0            5000
 normal    T

yes? DEF AXIS/X=123:223:5  newx
yes? DEF AXIS/Y=-20:20:5 newy

yes? LET outvar = temp[gx=newx,gy=newy]
yes? SHOW GRID outvar
 
yes? show grid outvar
    GRID (G002)
 name       axis    # pts   start        end
 NEWX      X          21 r   123          223
 NEWY      Y           9 r   -20          20
 ZAXLEVITR DEPTH (m)  20 i-  0            5000
 normal    T
 

-------------------------------------------
Now, an example of how to use scat2gridgauss.  See how the
variables LON and LAT are just 1-D lists of values, which  become
inputs to the function, and X[GX=XAX], Y[GY=YAX] are the
coordinates of the axes we want the result to be on.

yes? SET DATA scat.nc
yes? SHOW data

     currently SET data sets:
    1> /home/shiva5/data/island_rainfall.cdf  (default)
 name     title               I         J     K     L
 LAT      Latitude            1:226     ...   ...   ...
 LON      Longitude           1:226     ...   ...   ...
 RAIN     Island rainfall     1:226     ...   ...   1:840
 
yes? define axis/x=130:240:2.5 xax
yes? define axis/y=-35:20:2.5 yax

yes? let raingrid = scat2gridgauss_xy (lat, lon, rain,
 x[gx=xax], y[gy=yax],  4, 2, 4, 2)
yes? stat sgrid
 
             SCAT2GRIDGAUSS_XY (LON, LAT, RAIN,  X[GX=XAX], Y[GY=YAX],  4, 2, 4,2)
             X: 128.8 to 241.2
             Y: -36.2 to 21.2
             Z:  N/A
             TIME: 31-DEC-1970 12:00 to 19-APR-1973 12:00
             DATA SET: /home/data/island_rainfall.cdf
 
 Total # of data points: 869400 (45*23*1*840)
 # flagged as bad  data: 386543
 Minimum value: 0
 Maximum value: 380.4
 Mean    value: 5.8762 (unweighted average)
 Standard deviation: 12.64

Hemerson Tonin wrote:

Hi ferreter's

Do external functions (scat2gridgauss_xy and scat2gridlaplace)
interpolations from irregular axes to regular ones?

Because I have tryed with regular axis for longitude and irregular one for
latitude ( time and depth axes will not changed ) and, it has given me the
message:

              Bailing out of external function "scat2gridgauss_xy":
                       Input scattered x, y have different # of
points       4      11
              **ERROR: : error in external function

Well, suggestions to do this will be welcome !

Thank's

Tonin, H. E.
School of Chemistry, Physics and Earth Sciences
Flinders University of South Australia

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