Re: question about convert ECMWFds117.0 data into fix grid data

There are no single built-in functions that
will do what you want. I would suggest that the following
be built into a function.

The following is a sketch of what you need to do.

     f = addfile (" ....grb" , "r")

; spherical harmonic data
; the following will take a lot of memory because
; it is treating large array. If memory problems,
; loop over time and level

     x = f->X

     dimx = dimsizes( x )
     ntim = dimx( 0 )
     klev = dimx( 1 )
     nlat = 160
     mlon = 320

     temp = new ( (/ntim,klev,nlat,mlon/) , typeof(x) )
     delete(temp@_FillValue)

     a = x(:,:,0,:,:) ; real
     b = x(:,:,1,:,:) ; imaginary

     shsgc(a,b,temp) ; temp is on (160,320) gaussian grid

    ;use vinth2p to interpolate to pressure
    ; does not do a good job below the surface.

     temp_p = vinth2p(...) ; still on (160,320) gaussian grid

    ;interpolate to 0.5 grid
     NLAT = 360 ; 0.5
     NLAT = 720
     TEMP = g2fsh_Wrap(temp_p, (/NLAT,MLON/) )
or
     TEMP = linint2_Wrap(lon,lat,temp_p, True, LON,LAT,opt)
     printVarSummary( TEMP )

> In the same vein as Yinghui's question, I'd like to convert
> ERA40 ds117.2 (T159, native model levels) on a regular
> lat/lon/pressure grid, but at its full resolution (~0.5dg)
> instead of 2.5dg as in ds118.0. Are there existing NCL routines
> to do this operation?
>
> Thanks in advance,
>
> Francois
>
> On Wed, Jul 19, 2006 at 05:15:04PM -0600, Dennis Shea wrote:
>>> I am trying to convert the ECMWFds117.0 data to 2.5*2.5 degree data.
>>> I am using the NCL g2fsh function to do that. The problem is that
>>> ECMWFds117.0 data is in reduced N80 Gaussian grid, which need
>>> to be converted to regular N80 Gaussian grid first, I guess. Does anyone
>>> know a way to convert the reduced N80 Gaussian grid to regular N80
>>> Gaussian grid in NCL?
>>> _______________________________________________
>>
>> [1] NCL *automatically* converts the data to a
>> regular N80 Gaussian grid. NCL uses an ECMWF
>> supplied code to do this.
>>
>> [2] Why not use ds118.0? It is already on a 2.5 x 2.5 grid
>>
>> http://dss.ucar.edu/datasets/ds118.0/
>> ERA40 2.5 Degree Latitude-Longitude Gridded Surface
>> and Single Level Analysis from ECMWF
>>
>> [3] If you want to convert from GRIB to netCDF use NCL's
>> ncl_grib2nc ... a command line operator.
>>
>> http://www.ncl.ucar.edu/Document/Tools/ncl_grib2nc.shtml
>>
>> good luck
>>
>> _______________________________________________
>> ncl-talk mailing list
>> ncl-talk_at_ucar.edu
>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>
>
_______________________________________________
ncl-talk mailing list
ncl-talk_at_ucar.edu
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Jul 19 2006 - 20:46:19 MDT