Non-stationary extreme value models to account for the intensification of extreme typhoon waves

In the context of climate change, accurately calculating the return period values of typhoon waves is crucial for marine engineering. This study analyzes the long-term trends in the standard deviation of extreme typhoon wave height series, indicating an increased degree of dispersion (i.e., greater variability in significant wave heights). Current non-stationary methodologies for calculating return values of significant wave heights fail to account for changes in the degree of dispersion, thereby significantly affecting the accuracy of design wave heights. In response, this study proposes a new Poisson-Gumbel distribution model that incorporates changes in the degree of dispersion parameter (i.e., standard deviation) and systematically analyzes its validity. Using the South China coast as a case study, the application of a new method predicts that the 100-year return period significant wave height in the region will increase by 2.67 m over the next 100 years, whereas previous methods predicted an increase ranging from 0.3 to 1.5 m. The model presented in this paper calculates a 100-year return value of significant wave height that shows a 28.04% increase compared to the current value, leading to a corresponding 109.9% increase in the tetrapod weight for port breakwaters. • Reveals that increasing dispersion significantly impacts return values of significant wave heights. • Proposes a new non-stationary Poisson-Gumbel compound distribution model for extreme typhoon waves. • Incorporates dispersion trends and typhoon frequency into return value calculations. • The model predicts a 2.67 m (28.04%) increase in 100-year wave height and a 109.9% rise in tetrapod weight.