Subject: k-spectra and fw-spectra data
Date: Sat, 21 Jan 2006 15:29:42 -0800
From: Thierry R Michel <trm@jpl.nasa.gov>
Organization: JPL-334
To: Dimitris Menemenlis <menemenlis@sbcglobal.net>


===================Wavenumber Spectra===========================
The k-spectra data directory is:
c15:/u/trm/Ecco_wrk/K_ham1d  ("1d" is for the hole fill algorithm
                              "ham" is for Ham0.5 weighting)

Example figure: k_spect_ham05_all.ps or .pdf

Binary data 513*2: the first record (513f) is the spectrum
density (cm2*Km) from 0-wavenumber (not included in fig)
to 512/(1024*5.75Km).  The second record is the wavenumber
(0:512*1/(1024*5.75Km)). So, any integral of this function
gives the variance of the SSH residual in the included scales.


===================Wavenumber/Frequency Spectra===========================

The wk-spectra data directory is:
c15:/u/trm/Ecco_wrk/KW_tap2d  ("2d" is for the hole fill algorithm
                               "ham" is Ham0.5 window in space and time)


Here the dimension of the frequency axis depends on the length of the
data record and is in the file name "cube_cube18_513x178f.dat"
(178 fortran records of length 513*r4) where 178 = NF

With
  ds = 5750.0e-3Km; 
  dt =  9.94303days;

the wavenumber abcissa is again (0:512)*1/(1024*ds);
the frequency abcissa is (0:NF-1)*1/(2*(NF-1)*dt);

The Spectrum density unit is cm2*Km*Day...

You will notice that the density is too noisy for contour
plots.  I thik a color coded image would do... with
"smoothed" contours on top !?