readme.txt SeaState 2015, Fixed Riegl Lidar Wave Data BWB, Mar 01, 2016 This directory contains 10-minute summary wave statistics and spectra computed from the fixed-beam Riegl lidar deployed on the bow tower of R/V Knorr for the Davis Strait and HiWinGS 2013 cruises. Results are on a 10-minute timebase. Nothing is computed for hours when data is missing or loss is 100%. ------------------------------------------------------------------------------------------ Variance spectra aree computed 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). Spectra are only computed for 10-min intervals where data loss is < 50%. Smoothed spectral data is saved as an Nx241 array in SeaState_2015_WaveSpectra.txt SeaState_2015_WaveSpectra.mat and with the following format: columns variable description 1 Decimal Day-of-Year timestamp 2:79 Freq Frequency bin array, Hz 80:157 Df Delta F for bins, Hz 158:235 Sw Binned spectral data, m^2/Hz See wave_specreader_SeaState_2015.m for code to read & plot from this file. Raw 10 minute variance spectra are also saved in daily files 'WvSpecRaw_xxx.mat' in the Spectra directory. Data format is a single 144 x 2051 array. Column 1 is decimal day-of-year, columns 2:1026 frequency array, and columns 1027:2051 spectral data. ------------------------------------------------------------------------------------------ 10 minute wave stats file SeaState_2015_WaveStats_Hw_hdr.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 18 Fw estimated separation frequency for wind sea 19 windspeed from NOAA sonic on bow tower, m/s 20 Hw estimated significant wave ht for wind sea Hour average stats are in SeaState_2015_WaveStats_Hw_hr_hdr.txt with the same column format. Note that the 'missing' variable in this file is the average number of missing in each 10 min interval. Multiply by 6 to get the total number of missing over the whole hour. 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 >50%. 4) Fw and Hw are computed with assumption that U10n/Cp > 0.8 for wind sea equilibrium, thus Cp = U10n/0.8 and Fw = 0.8*g / (2*pi*U10n). Hw is computed from the binned spectra (Swv), frequency bins (fwv) and delta frequency (dfw) as: ii = find(fwv>=Fw,1,'first'); % start bin for integrations hw = 4*sqrt(sum(Swv(ii:end).*dfw(ii:end))); ------------------------------------------------------------------------------------------ The directory Corrected_Sfc_Ht contains hourly files of ocean surface displacement. These are lidar data corrected for heave, pitch and roll, with missing values filled by interpolated to exactly 10 Hz. The first column is timestamp decimal day of year. The second column is surface displacement from the hourly mean in meters. One caution: for hours with a large amount of data loss the interpolation generates garbage. Consult the 'missing' variable in SeaState_2015_WaveStats_Hw_hdr.txt or SeaState_2015_WaveStats_Hw_hr_hdr.txt as a filter for bad time periods. In many cases, as long as data loss is less than 50% and evenly distributed (i.e. approximately every other data point lost) the interpolated result is reliable. Also note that sea spray thrown up at the ship's bow can cause large positive spikes in the surface displacement timeseries when the ship is underway, and these have not been effectively removed from the data. 10 Hz d