Ryota Wada, University of Tokyo


A simple spatial extreme model for ocean waves


We seek to characterize the behavior of extreme waves in the ocean for engineering purpose. For extrapolation to long return periods, estimation using data for around 100 years from a single location will incur large uncertainties. In many cases, the estimation is based on data set from a single location and the spatial dependency structure of the largest waves is not considered. In the presentation, a simple spatial statistical model for extreme value estimation of significant wave height under storms, known as STM-E, will be explained. The STM-E model was developed to characterize extreme waves offshore Japan, dominated by tropical cyclones. The method relies on the estimation of two distributions from a sample of data, namely the distribution of spatio-temporal maximum (STM) and the exposure (E). The STM-E estimate provides a parsimonious spatially-smooth distribution of extreme waves, with smaller uncertainties per location compared to estimates using data from a single location. In addition, STM and exposure provide a good understanding of extreme phenomena. We also discuss the estimated characteristics of extreme wave environments in this region. Utilization of this approach can be extended to various extreme value estimation with spatial dependency structure.


Tuesday September 10th, 11:15


Rossbysalen C609, Arrhenius laboratory, Svante Arrhenius väg 16C, 6th floor