R.J. Sobey, S.A. Hughes r Coastal Engineering 36 (1999) 1736
35
unsuitable. Pressure measurement at near-surface elevations under waves would require
special attention.
8. Conclusions
Traditional directional analysis of PUV data results in linearized estimates of
directional wave and velocity spectra which typically do not include steady currents.
Attempts to reconstitute velocity time series at any other location in the water column
result in linear estimates with decreasing veracity approaching the free surface. The
single advantage of linear directional spectral analysis is the compactness of the output
which corresponds to loss of information. However, this compactness could become a
liability in some applications.
A new method has been presented for interpreting traces recorded by conventional
PUV instruments under irregular multidirectional waves. This time-domain method
provides estimates of sea surface elevation and wave kinematics throughout the water
column without compromising the essential nonlinear character of the waves. The key to
the formulation is retention of the fully nonlinear free surface boundary conditions
which are used in seeking a nonlinear least-squares solution in a narrow time window.
Also included is the nonlinear interaction between the waves and a steady and
depth-uniform current. The analysis method provides nearly exact results for theoretical
nonlinear steady waves. Application to field data requires overspecification in the
solution to allow for instrument error bands. Nevertheless, credible solutions were
obtained using PUV measurement from two field locations with very energetic wave and
current conditions.
Results from the locally nonlinear analysis differ substantially from traditional
directional spectra. The method provides time series of sea surface elevation h, radial
frequency v , directional wave numbers k1 and k 2 , and Fourier coefficients, A j. These
local solution parameters provide predictions of corresponding time series for any
kinematic quantity at any position in the water column. A complete field solution is
available at each time interval.
The locally nonlinear PUV method provides a distinct improvement for analyzing
practical problems in which individual wave characteristics are critical. For example,
wave loads on offshore structures depend on wave height, fluid velocity, fluid accelera-
nonlinear characteristics inherent in the measured data is critical for credible load
determination.
The locally nonlinear analysis is fairly computationally intensive, and it creates
additional information rather than condensing the measurements. Therefore, this method
is probably not well-suited for routine analysis and archiving of PUV records obtained
over a lengthy deployment. However, because the method does not rely on the entire
time series, it is possible to perform selective analysis of individual waves or large wave
groups within the time series that may be of particular interest. Examples include
so-called `freak waves' and even tsunami waves. Also, the local nature of the nonlinear
analysis method may find application providing near real-time wave estimates from