A Kernel Density Estimation Approach and Statistical Generalized Additive Model of Western North Pacific Typhoon Activities

Abstract

This paper presents a development of a statistical model of typhoon genesis, tracks based on kernel density estimation and a generalized additive model (GAM). Modeling of typhoon activity is ultimately beneficial to the people living in coastal zones, insurance/re-insurance companies, policy, planning and decision departments. A 50-year record (1972–2021) of typhoon track observations from the International Best Track Archive Climate Stewardship have been used to observe the distribution of typhoon genesis by kernel density estimation. The tracks are simulated through the development of a GAM. It reproduces the observation well. A distance calculation approach between observed and simulated tracks’ landfall have been used to validate the model and the model shows a very good skill (approximately 75%).

Keywords

• kernel density estimation
• generalized additive model
• typhoon
• Western North Pacific