Abstract
Discus N denotes a single buoy employed during the SWADE experiment, equipped to record wave amplitude and wind speed time series at a rate of 1 Hz. Over the course of approximately 4.5 days, two clear-cut examples of sea response to wind activity took place. It is easy to verify that the spectral content of the time series is changing. Wavelet analysis (WA) is a powerful tool for analyzing such nonstationary series. The paper illustrates the use of this technique to characterize the observed wave response in a quantitative manner and to compare this response to simultaneously measured wind state data. For reasons that will be reviewed, unlike Fourier analysis, a WA requires “fine-tuning” of the basis functions to fit the problem under consideration. Within geophysical applications it has become common to utilize the “Morlet” wavelet because of its strong resemblance to well-known spectrogram analysis techniques. However, it will be seen that a relatively new technique known as the discrete wavelet packet transform is in principle especially well suited to optimal time-frequency localizations that are useful in analyzing nonstationary processes.
