Readme.txt
SeaState 2015, Fixed Riegl Lidar Wave Data
BWB, Feb 5, 2016

This directory contains summary wave statistics and spectra computed from the fixed-beam Riegl
lidar deployed on the bow tower of R/V Sikuliaq for the SeaState 2015 cruise.

Results are on a 10-minute timebase.  Nothing is computed for hours when data loss is 100%.

------------------------------------------------------------------------------------------
Variance spectra were computed 2 ways:

    1) from the autocovariance coefficients of the motion-corrected surface displacement
        timeseries.  This method is tolerant of missing data (NaNs).  A Hamming window is
        applied to the raw timeseries but it is not high-pass filtered.  Low-frequency drift
        and high frequency artifacts are present;
    2) computed with the psd function from the motion-corrected surface displacement timeseries
        with missing data interpolation (pchip).  The time series is high-pass filtered
        (0.04 Hz stopband, 0.05 Hz passband).
        
    The spectral peak from method 1 usually corresponds well to that from method 2.  The
    method 2 result is recommended for general use.
    
    Spectra are only computed for 10-min intervals where data loss is < 15%.

Smoothed spectral data is saved as a 5040x245 array in SeaState_2015_WaveSpectra.txt 
with the following format:

columns   Description
1         jd, Decimal Day-of-Year timestamp
2:51      F1, Frequency bin array, method 1, Hz
52:101    S1, Spectral data, method 1, m^2/Hz
102:149   F2, Frequency bin array, method 2, Hz
150:197   Df2, Delta F for frequency bins, Hz
198:245   S2, Spectral data, method 2, m^2.Hz

See wave_specreader_SeaState_2015.m for code to read / plot from this file.


------------------------------------------------------------------------------------------
10 minute wave statistic file SeaState_2015_WaveStats.txt contains missing data count and
wave stats in the following format:

column    Description
1         jd Decimal Day-of-Year timestamp
2         # missing points in each 10 min segment
3         Hs from crude motion correction (-range + heave), m
4         Hs_cf from CF motion correction method, m
5         Hs_spec from spectral moment (4*sqrt(m0)), m
6         Tz mean zero crossing period from spectral moments (sqrt(m0/m2))
7         Te energy period from spectral moments (mn1/m0)
8         Tpc approximate wave peak period from spectral moments (mn2.*m1)./(m0^2)
9         m0 spectral moment (sum(S.*df))
10        m1 spectral moment (sum(f.*S.*df))
11        m2 spectral moment (sum((f.^2).*S.*df))
12        m3 spectral moment (sum((f.^3).*S.*df))
13        m4 spectral moment (sum((f.^2).*S.*df))
14        mn1 spectral moment (sum((f.^-1).*S.*df))
15        mn2 spectral moment (sum((f.^-2).*S.*df))
16        Fp frequency of highest spectral peak in 0.03-0.3 Hz band, Hz
17        Tp peak period = (1/Fp), sec

Notes:
1) Variable 3, Hs, is computed as 4*std(h), where h is the crudely corrected surface
    displacement timeseries ( -lidar_range + heave ).  Heave is derived from the NOAA
    flux system motion correction.  Computed for every interval with data.
2) Variable 4, Hs_cf, is computed as 4*std(hc), where hc is corrected for pitch, roll
    and mount angle in addition to heave.  Computed for every interval with data.
3) Variable 5, Hs_spec, is computed as 4*sqrt(m0).  Probably the best value for times
    when spectra are computed.  Not computed when missing >15%.