Abstract
This study investigates the spatial and spatio-temporal prediction of significant wave height (SWH) and corresponding wave power analysis in the seas surrounding the Korean Peninsula using fixed rank kriging (FRK). Using both reanalysis data and observational data, the analysis addresses the non-stationary and non-Gaussian characteristics inherent in SWH data, and refines the SWH data to a finer spatial resolution for conducting a more detailed wave power analysis. In this study, two different modeling scenarios were investigated: seasonal spatial modeling and daily spatio-temporal modeling. The first scenario explores seasonal patterns of SWH, identifying the offshore waters along the eastern coast of the Korean Peninsula in winter as the region with the highest wave power potential. The second scenario captures daily variations through a moving window approach, incorporating temporal dependencies for more accurate short-term predictions. To handle the non-Gaussian nature of the data, we considered Gaussian, Inverse Gaussian, and Gamma distributions, with the Gamma distribution consistently yielding the best predictive performance across all scenarios. The FRK framework demonstrated effective predictive capability over large spatio-temporal scales. The results provide practical insights into optimal site selection for wave power generation and highlight the feasibility of using wave energy as a renewable energy source in South Korea