and the approach to burial are indicated in the SSC signal by rising
background or change in the sensor offset. Complete burial is
indicated by a significant change in sensor offset. Subjective
estimates of when bio-fouling or burial became significant and
affected data were discarded from analysis.
(5) Processed time series data were output to *.tsc files identical in
format to the extracted *.ts format.
(6) Burst-averaged summary statistics files (*.sts) were generated by
taking the mean and variance of the processed data for each 2048-
point burst.
c. Matlab routine plotbursts_*.m was used to plot time series to conduct
final QC on data, remove spikes and velocity ambiguities, calculate and
plot autospectra, cross-spectra, and to calculate a correlation score on
velocity data. The correlation score expresses the percentage of the
record in which velocity correlations are above a threshold. In this case,
a threshold of 70 percent was the a basis for selecting useful velocity
data.
Data post-processing was performed on the velocity (E, N, U) data to remove
poor quality or erroneous data. Poor quality data is typically a result of
environmental conditions, which cause poor acoustic signal return and low signal
correlation. Instrument motion problems can also cause velocity ambiguities that
are out of range. Time series plots of the measured burst-averaged parameters
subjected to post-processing are provided in Figures D10 to D108.
Current transect data were initially processed with ADP manufacturer's
software to correct for vessel movement using ADP bottom track or DGPS-
derived vessel velocity. Data quality filtering was performed to remove velocity
measurements with low signal-to-noise ratios. The resultant speed and direction
data were horizontally smoothed using a 5- or 7-point Gaussian filter (profile
averaging) and vertically smoothed using a 3-point Gaussian filter (cell
averaging). This filtering helps to minimize some of the uncorrected high-
frequency velocity error resulting from vessel heave, pitch, roll and rapid turning.
Depths measured simultaneously with current data were corrected using a water-
surface elevation time series to produce bottom elevations referenced to datum.
Spikes in depth data and noise in the digital depth data were cross-referenced to
paper echosounder records and erroneous data were either smoothed or deleted.
Current profile speed and direction were merged with bottom elevation data
to produce cross-sectional plots of speed and direction relative to distance along
the planned transect line. Plots of cross-sectional ADP transect data are shown in
Figures D39 to D100.
Vertical speed profiles were depth-averaged from 1 m below the water
surface to 0.5 m above bottom. Current direction was first multiplied by the
current speed and then averaged to obtain a speed-weighted depth-averaged
current direction. Depth-averaged speed and direction vectors were then scaled
and plotted as plan-view vectors overlaid on a chart of the study area (Examples
are shown in Figures D8 and D9).
D18
Appendix D
Field Data Collection
Previous Page
