[gradsusr] Fw: EOF and spectral analysis on GrADS

[Jose]

Dear Tamer Nada

Thank you by your explanation and suggestions. They were clear.

By the way do you know where can I found the fortran subroutines from lepak library you commented? 


Thanks in advance

Enviado via iPad

Em 26/01/2013, às 16:53, Tamer Nada <tamer540 at gmail.com> escreveu:

> Dear :- Jose Augusto Paixão Veiga
> 
> In Fortran :-
> 
> You must know there is about 3 methods ( algorithms )  to calculate eigen Vector and eigen Value  each method some of it efficient but just compute only first Eigen value and first Eigen vector by suggest value and using iteration methods to compute it until reach to the difference between old value from last iteration to new value equal small fraction you can determine it . another to methods using "SVD "  and this method its divided by to relative to the shap of data matrix   and some of it efficient but slow and some fast and not efficient 
> 
> I suggest to you to use lepak library in fortran ( its easy to call subroutine ) or using high level language or statistical or climatic package like NCL or R or Matlab 
> 
> On Sat, Jan 26, 2013 at 9:56 PM, Jose Augusto Paixão Veiga <veiga.uea at gmail.com> wrote:
>> Dear EOF expertise,
>> 
>> one of you would have a material given instruction (if possible step by step) on how to calculate EOF?
>> 
>> I would like to do a fortran program, from the initial point (for example: 1-read a 3-d matrix data, 2-calculate its inversion, 3-multiply it by its ....), to calculate EOF for climate study purposes.
>> 
>> A flux diagram will be very appreciated. 
>> 
>> Thanks in advance.   
>> 
>> Augusto
>> 
>> 
>> On Jan 26, 2013, at 3:41 PM, Tamer Nada <tamer540 at gmail.com> wrote:
>> 
>>> on Grads No, but you can use NCL  or R to compute EOF; and Only R  or Matlab to Compute CCA ,But also you can Use NCL to compute "SVD" and after that you can compute CCA ,actually NCL have function to compute CCA ,But not efficient because  its use old IBM Fortran subroutine ,so its not work fine with big array of data and may be hang in calculation ,so i prefer to use SVD function and after that you can compute CCA    
>>> 
>>> On Sat, Jan 26, 2013 at 9:58 AM, ugo okoro <uknac23 at yahoo.com> wrote:
>>>> 
>>>> Dear all,
>>>> 
>>>> Please any info on how to run EOF and spectral analysis (frequency-time) on GrADS?
>>>> 
>>>> Anticipating prompt response.
>>>> 
>>>> Thanks.
>>>> 
>>>> Ugo Okoro.
>>>> CMSR-IAP
>>>> CAS, Beijing.
>>>> 
>>>> _______________________________________________
>>>> gradsusr mailing list
>>>> gradsusr at gradsusr.org
>>>> http://gradsusr.org/mailman/listinfo/gradsusr
>>> 
>>> 
>>> 
>>> -- 
>>> =====================
>>> Tamer Ali Nada 
>>> Research Worker
>>> Cairo Numerical Weather Prediction Center { CNWP}
>>> Egyption Meteorological Authority   { EMA }
>>> Cairo , Egypt
>>> ============================
>>> tamer540 at yahoo.com
>>> tamer540 at gmail.com
>>> tamer.nada at noaa.gov
>>> _______________________________________________
>>> gradsusr mailing list
>>> gradsusr at gradsusr.org
>>> http://gradsusr.org/mailman/listinfo/gradsusr
>> 
>> 
>> _______________________________________________
>> gradsusr mailing list
>> gradsusr at gradsusr.org
>> http://gradsusr.org/mailman/listinfo/gradsusr
> 
> 
> 
> -- 
> =====================
> Tamer Ali Nada 
> Research Worker
> Cairo Numerical Weather Prediction Center { CNWP}
> Egyption Meteorological Authority   { EMA }
> Cairo , Egypt
> ============================
> tamer540 at yahoo.com
> tamer540 at gmail.com
> tamer.nada at noaa.gov
> _______________________________________________
> gradsusr mailing list
> gradsusr at gradsusr.org
> http://gradsusr.org/mailman/listinfo/gradsusr
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://gradsusr.org/pipermail/gradsusr/attachments/20130126/ff8d4bdf/attachment-0003.html