Abstract
Reasonable assessment of extreme tropical cyclone waves is of great significance to the development of marine renewable energy. However, the significant spatiotemporal variation of these waves poses a challenge to assessment. To cope with temporal variations, the peak over threshold method with flexible window is used to extract extreme samples, which makes full use of tropical cyclone samples. To cope with spatial variations, a regional extrapolation model is proposed by analysing the three-dimensional interrelation of the at-site probability distribution. In this model, a spatial control law is constructed based on site and wave characteristics to govern the local response of the probability distribution, and the valid sample information is added by trading space for time. By using this model, return significant wave heights with reduced uncertainty can be derived at sites with insufficient samples. In addition, the spatial distribution of return significant wave height for different return periods in the study region can also be derived, revealing the spatiotemporal characteristic of extreme tropical cyclone waves. These results provide important support for wave disaster prevention and farm deployment of marine renewable energy industry.