Inherent uncertainties in numerical modeling of hurricane intensity

It is widely accepted that current operational models have little or no skill with forecasting tropical cyclone intensity. A major reason that is often mentioned is the lack of a coupled ocean in most current systems. Other reasons include lack of resolution and inability to model the finer-scale physical processes, such as the surface boundary layer feedbacks. Implicit is the assumption that the model improvements will logically lead to substantial improvements in intensity forecasting. In this study we report on the inherent uncertainties in utilizing a numerical model to either predict or simulate tropical cyclone intensity. We take an idealized hurricane simulation using our TCM3 model and examine the effects of model resolution and numerics, cloud and physical parameterizations, and limitations arising from the need to accurately model mesoscale feature of the cyclone environment. Our preliminary findings, which will be updated for the meeting, indicate that inherent uncertainty of up to 30-40% of the local maximum potential intensity is likely to be experienced.

This study forms part of a larger effort examining the predictability of hurricane tracks and structure. The context within the larger effort will be reported.