An Environmentally Forced Tropical Cyclone Hazard Model

Abstract

Abstract A new statistical‐dynamical model is developed for estimating the long‐term hazard of rare, high impact tropical cyclones events globally. There are three components representing the complete storm lifetime: an environmental index‐based genesis model, a beta‐advection track model, and an autoregressive intensity model. All three components depend upon the local environmental conditions, including potential intensity, relative sea surface temperature, 850 and 250 hPa steering flow, deep‐layer mean vertical shear, 850 hPa vorticity, and midlevel relative humidity. The hazard model, using 400 realizations of a 32 year period (approximately 3,000 storms per realization), captures many aspects of tropical cyclone statistics, such as genesis and track density distribution. Of particular note, it simulates the observed number of rapidly intensifying storms, a challenging issue in tropical cyclone modeling and prediction. Using the return period curve of landfall intensity as a measure of local tropical cyclone hazard, the model reasonably simulates the hazard in the western north Pacific (coastal regions of the Philippines, China, Taiwan, and Japan) and the Caribbean islands. In other regions, the observed return period curve can be captured after a local landfall frequency adjustment that forces the total number of landfalls to be the same as that observed while allowing the model to freely simulate the distribution of intensities at landfall.

Keywords

• tropical cyclone
• hazard assessment
• risk assessment
• statistical‐dynamical downscaling
• tropical cyclone climatology