be +3.5 times the standard deviation about the mean. Data points outside the data

limits are discarded, and the mean value is recalculated for the remaining data set

within the sliding window. Smaller values of Alpha sometimes are used for

ADVs 1 and 2 because these signals contain many glitches that result from the

presence of air bubbles in shallow water. However, good judgement is required

in the selection of Alpha, because valid parts of the signal may be removed. The

remaining set of valid data points is sorted into increasing order by time, and

linear interpolation is used to replace points removed by the program. Therefore,

the length of the filtered time series is the same as the original time series, and

the 38 channels of current data remain synchronized. Normally, only ADVs 1

through 4 need to have glitches removed from the velocity signal, since these are

the sensors that are located in shallow water depths.

At this step in the analysis, the batch file pauses allowing the user to

superimpose the filtered time signal onto the original time series, which contains

spikes caused from the entrained air bubbles. This is integrated into the GEDAP

batch routine and is plotted using GPLOT, as shown in Figure 46. This figure

shows the results of filtering the y-axis (longshore) time series for ADV 2. These

results were obtained using a value of Alpha = 3.5, which works well if the true

signal has a Gaussian distribution. It sometimes requires several iterations to

determine the optimum value of Alpha. However, it should be mentioned that in

some cases, the number of spikes are so numerous that the signal cannot be

properly filtered to obtain a time series free of spikes. If this is the case, the time

series is discarded. A more complete description of this filtering routine is

available in the GEDAP on-line documentation.

After the signals in shallow water have been filtered, the program STAT1 is

used to calculate the minimum, mean, maximum, and standard deviation of the

velocity signals for both the x- and y-axes. The mean value obtained from the y-

axis of the 10 ADVs defines the cross-shore distribution of the longshore current.

Finally, the GEDAP program PLUCK is used to collect the values of the four

statistical parameters calculated using STAT1 from the header of each of the

output files. The final result is an ASCII file consisting of six columns of data.

The first two columns identify the channel number and ADV number and axis.

The last four columns contain the values of the four statistical parameters

discussed above for all 38 ADV channels.

As mentioned previously, as many as 12 transects have been measured

during a 2 hr segment of an experiment. In contrast to the GEDAP batch file

used for postprocessing the wave data, the batch file processes only one transect

of current data, because several iterations may be necessary to determine an

optimum value of Alpha. One transect of current data can be analyzed in

approximately 15 min. Therefore, approximately 3 hr would be usually be

required to process12 transects of current data.

Examples of the cross-shore distribution of the measured mean longshore

current during a regular wave and irregular wave experiment are shown in

Figures 66 and 68, Chapter 9.

83

Chapter 8

Sensor Calibration, Measurement, and Data Analysis

Integrated Publishing, Inc. |