Analyses of sea clutter measured at Taipei Harbour were carried out. Statistical models commonly used by researchers for sea clutter were used to fit measured data. These include the Rayleigh, the Weibull, the lognormal, and the so-called (compound) K- distributions. It is shown that the lognormal distribution results in the best fits among all possible candidates. Possible sources for the misfits of the K-distribution suggested by many researchers were also discussed.
Sea clutter is the backscattered electromagnetic waves of radar from the sea surface. While they may be considered annoying for the most of time, sea clutters may contain information of the structure of the water surface and are, therefore, considered by oceanographers to be rather useful. It has been shown by many researchers that, valuable information concerning both the wave- and wind-fields can be derived from it (Gommenginger, 1997, see also Gommenginger et al., 2000; Robinson et al., 2000; as well as Lentine, 2006). Young et al. (1985; Ziemer, 1987; Gangeskar; 2000) have shown that, useful information concerning the wave fields can be extracted from radar image sequences. Using wave heights obtained from the so-called wavenumber-frequency spectrum, Nieto Borge & Guedes Soares (2000; Izquierdo et al., 2004, 2005; Hessner et al., 2006) clearly show that these are in good agreements with on site measurement. Furthermore, it have also been shown that surface current speeds and direction (Senet, 1996), as well as water depth (Outzen, 1998), and information about the wind fields above the water surface (Hatten, 1998; Dankert, 2003) can be inferred from radar images. Statistical analyses of sea clutter started with the intension of its suppression so that objects other than waves on the sea surface can be detected easily (Ward et al., 1990a, b; Wetzel, 1990). Most commonly used statistical models are, the Rayleigh, Weibull, lognormal, and the so-called K-distributions.