Re: linear interpolation

Why not just use the netcdf toolbox in Matlab and save your files as .nc
files that can be easily opened in NCL?

http://mexcdf.sourceforge.net/

Dave Small

On 23/09/09 4:02 PM, "Lei Meng" <dream916_at_gmail.com> wrote:

> Hi Dennis,
>    Thanks for your suggestions and it works fine. Except that the longitudes
> from the two nc files do not match. The longitudes from one file are ranged
> from -179.5 to +179.5 and the ones from the other file are from 0 to 357. When
> I did the linear interpolation, only the values for longitudes 0 to 179.0
> retain and there are no values for longitudes from -179.5 to 0. Any
> suggestions?
>   Thank you for your help!
>   Lei
>
>
> On Wed, Sep 23, 2009 at 2:13 PM, Dennis Shea <shea_at_ucar.edu> wrote:
>> To further what DaveB wrote:
>>
>> You did not state if the input grid was on hybrid levels or pressure levels.
>> There are slightly different approaches that may be used.
>>
>> Generally,  the vertical interpolation should be done first:
>>
>>    X(12, 60,180,360)  ===> Y(12, 28,180,360)
>>
>> Then, perform the horizontal interpolation:
>>
>>   Y(12, 28,180,360)  ===> Z(12, 28, 64,128)
>>   ===========
>> NCL's Applications page is a good place to start:
>>
>>          http://www.ncl.ucar.edu/Applications/
>>
>> ==================== VERTICAL INTERPOLATION ===============
>> Under the "Data Analysis", Click:  "Vertical Interpolation"
>>
>>         http://www.ncl.ucar.edu/Applications/vert_interp.shtml
>>          If your input grid is on hybrid levels, Example 3 is closest to
>> what you want to do.
>>
>> If the input is on pressure levels, the function "int2p"
>>
>>          http://www.ncl.ucar.edu/Document/Functions/Built-in/int2p.shtml  
>> or, better, "int2p_Wrap", because it returns meta data
>>
>>          
>> http://www.ncl.ucar.edu/Document/Functions/Contributed/int2p_Wrap.shtml  
>> See Example 5
>>
>>  yTemp = *int2p_Wrap*(p_in,  x(time|:,lat|:,lon|:,pin|:), pout, linlog)
>>  printVarSummary(yTemp)
>>
>>  YY = yTemp(time|:,iLev|:,LAT|:,LON|:)   ; reorder for horizontal
>> interpolation
>>  delete(yTemp)
>>
>> use, linint2_Wrap, or f2gsh_Wrap or area_hitolores for the grid
>> interpolation.
>>
>>
>>
>>
>>
>> David Brown wrote:
>>> Hi Lei,
>>>
>>> First just a bit of terminology: you are talking about the dimensions  of
>>> the variable not the attributes (which are metadata such as units  or a
>>> descriptive long name for the variable). There are a number of  functions
>>> you could use. Assuming you have coordinate values for the  input and output
>>> lat and lon dimensions, and you are using version  5.1.0 of NCL, I would
>>> recommend the function area_hi2lores_Wrap.  However, you could also use
>>> linint2_Wrap, or if the output grid is  Gaussian ("T42") the function
>>> f2gsh_Wrap might be preferable.
>>>  -dave
>>>
>>>
>>> On Sep 22, 2009, at 1:05 PM, Lei Meng wrote:
>>>
>>>  
>>>> Hi,
>>>>  I have a nc file with attributes (Time, Lev, Lat, Lon) (e.g., 12,
>>>>  60,180,360) and I would like to convert it to a nc file with  attributes
>>>> (Time, iLev,iLat,iLon) (e.g., 12, 28, 64,128) using  linear interpolation.
>>>> Which functions should I use to interpolate  the three attributes (ilev,
>>>> ilat, ilon)? Thanks,
>>>>  Lei
>>>> _______________________________________________
>>>> ncl-talk mailing list
>>>> List instructions, subscriber options, unsubscribe:
>>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>>    
>>>
>>> _______________________________________________
>>> ncl-talk mailing list
>>> List instructions, subscriber options, unsubscribe:
>>> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>>>  
>>

_______________________________________________
ncl-talk mailing list
List instructions, subscriber options, unsubscribe:
http://mailman.ucar.edu/mailman/listinfo/ncl-talk
Received on Wed Sep 23 2009 - 14:21:13 MDT