8.3. WAVE DATA ANALYSIS
495
data. In other cases, backup instruments may have collected the same data
as the primary instrument, and there is no need to analyze the redundant
records provided the primary measurements were of high quality. Finally,
some data may have been collected that are not of immediate interest, but
instead intended for some type of future application.
Types of analyses range from simple to sophisticated, depending on the
particular experiment and the needs of the researcher. The most elementary analyses simply involve converting recorded data into a more useful
form (e.g., converting digital wave gauge voltages to equivalent sea surface
elevations expressed in engineering length units). More complicated analyses apply mathematical relationships and theories to convert the data from
one type of data into another type of data. Examples include representing measured sea surface elevations as probability distributions, converting
model bathymetry to maps showing depth contours, and integrating measured pressure distributions on a structure to obtain forces.
Time series data are commonly analyzed using time domain (statistical) techniques or frequency domain (spectral) techniques; and many good
textbooks have been written dealing with these analysis methods in splendid detail (e.g., Bendat and Piersol 1971). The following sections overview
the routine analyses performed on wave data collected in laboratory wave
flumes and basins. In most cases, the laboratory wave computer analyses
are very similar or identical to those used for field data, and the techniques
for implementing the analyses are well known. Consequently, the mathematical details of the analysis methods have been minimized here.
8.3.1 Data Editing
Wave data-editing is an optional process whereby recorded wave data are
manually or automatically manipulated for specific purposes. There are
several reasons for performing data editing:
1. Wave records may contain “glitches” or “spikes” that arise
from problems with the instrument or data acquisition, and
it is important to remove these errors in the data before
further analyses. Mansard and Funke (1988) described an
automatic technique for identifying glitches in wave time
series records and correcting the errors by linear interpolation between adjacent good points.
2. Some recorded wave data may contain an offset due to
the instrument or the analog-to-digital conversion, or there
may be a slowly varying trend due to instrument drift or
actual changes in water level. In any event, these offsets
495
data. In other cases, backup instruments may have collected the same data
as the primary instrument, and there is no need to analyze the redundant
records provided the primary measurements were of high quality. Finally,
some data may have been collected that are not of immediate interest, but
instead intended for some type of future application.
Types of analyses range from simple to sophisticated, depending on the
particular experiment and the needs of the researcher. The most elementary analyses simply involve converting recorded data into a more useful
form (e.g., converting digital wave gauge voltages to equivalent sea surface
elevations expressed in engineering length units). More complicated analyses apply mathematical relationships and theories to convert the data from
one type of data into another type of data. Examples include representing measured sea surface elevations as probability distributions, converting
model bathymetry to maps showing depth contours, and integrating measured pressure distributions on a structure to obtain forces.
Time series data are commonly analyzed using time domain (statistical) techniques or frequency domain (spectral) techniques; and many good
textbooks have been written dealing with these analysis methods in splendid detail (e.g., Bendat and Piersol 1971). The following sections overview
the routine analyses performed on wave data collected in laboratory wave
flumes and basins. In most cases, the laboratory wave computer analyses
are very similar or identical to those used for field data, and the techniques
for implementing the analyses are well known. Consequently, the mathematical details of the analysis methods have been minimized here.
8.3.1 Data Editing
Wave data-editing is an optional process whereby recorded wave data are
manually or automatically manipulated for specific purposes. There are
several reasons for performing data editing:
1. Wave records may contain “glitches” or “spikes” that arise
from problems with the instrument or data acquisition, and
it is important to remove these errors in the data before
further analyses. Mansard and Funke (1988) described an
automatic technique for identifying glitches in wave time
series records and correcting the errors by linear interpolation between adjacent good points.
2. Some recorded wave data may contain an offset due to
the instrument or the analog-to-digital conversion, or there
may be a slowly varying trend due to instrument drift or
actual changes in water level. In any event, these offsets