A Statistical Comparison of Small-Scale Features in Simulated Tropical Cyclones and High-Resolution Observational Data

Over the past few decades, there has been significant improvement in tropical cyclone track prediction but not much improvement in intensity prediction. There is evidence that small-scale processes in the inner core may play an important role in hurricane intensity but these processes, and what triggers them, are only partially understood. The features -- eyewall vorticity waves, mesovortices and hot towers -- are on the order of kilometers wide and rapidly varying. Hence high-resolution models are thought to be necessary, but it is not clear how well high-resolution features in simulated tropical cyclones match those in the real world. It is also unclear what resolutions are necessary to allow a simulation to accurately model the effects of these features on the large-scale state. This research uses WRF-based numerical weather prediction models to simulate tropical cyclones that have been observed to have periods of strong inner core activity. We compare statistical properties of high- resolution features in the simulation to the same statistical properties in observational data. Using that methodology, we analyze how these periods are represented in the model under different large-scale flow patterns and using different model resolutions and parameterizations.