近岸带波高与周期分布的核密度估计

使用双变量核密度估计方法描述近岸带波高和周期联合概率密度分布与波高、周期边缘密度分布。结果表明,核密度估计方法比通常使用的参数模式能更好地显示出具有多峰的波要素统计结构,核密度估计的波周期带宽系数能反映波浪谱的某些信息,尤其以波周期带宽和谱宽参量具有良好的线性关系。

KERNEL DENSITY ESTIMATES ON DISTRIBUTION OF WAVE HEIGHT AND PERIOD IN NEARESHORE ZONE

Statistical distribution of wave elements is important for understanding wave processes. At present, the distribution of wave period and that of wave height, or joint distribution of the both are described mainly by parametric distribution functions. For multi-peak waves that commonly seen in nearshore zones, a conventional approach would not work as it works only for single-peak cases. Other nonparametric estimates are therefore needed. Using wave trains of numerical simulation of Ochi-Hubble spectrum of two-peak and an artificial three-peak spectrum, kernel density estimate was tested for describing the joint distribution of wave height and period. The joint distribution of wave-period-and-height and corresponding marginal probability density were simulated with two-variable kernel density estimate for the case of nearshore coastal zone east of Guangdong, China. The results indicate that a multi-peak statistical structure could be described by kernel density estimate, but by conventional parametric models. Moreover, some features of wave spectrum were found well related to the bandwidth of wave period.