Are Rapidly Intensifying Tropical Cyclones Associated with Unique Vortex Characteristics?

Accurately predicting tropical cyclone (TC) intensity change is particularly challenging during rapid intensification (RI) events. Recent work has shown that operational TC intensity forecasts exhibit the largest errors when RI occurs in relatively favorable environmental conditions, suggesting an improved understanding of the vortex and convective characteristics associated with RI may aid TC intensity forecasts. This study explores whether TCs that undergo RI are associated with unique vortex characteristics compared to non-RI TCs. To accomplish this, TC vortex characteristics are assessed using the Tropical Cyclone Radar Archive of Doppler Analyses with Recentering (TC-RADAR) database, which contains approximately 1,200 airborne radar analyses from over 350 flights into TCs across a wide range of intensities. In this study, TCs are placed into one of three intensity change groups based on the future 12-h TC intensity change according to the best-track database: 1) Rapid intensification (RI), 2) slow intensification (SI), and 3) non-intensifying (NI). Because TC vortex characteristics are also related to the intensity of the storm, a normalization technique is used to identify the anomalous vortex characteristics of each TC, relative to TCs of similar intensities. Here, we will show that the anomalous azimuthally-averaged structure of RI TCs displays statistically significant differences from both SI and NI TCs. These differences are reflected in both the primary and overturning circulations of the storm. A bivariate parameter space consisting of metrics of 1) TC vortex favorability and 2) synoptic-scale environmental favorability for TC intensification is used to illustrate the multi-scale nature of TC intensity change. RI is shown to occur preferentially when both the environmental and vortex metrics are relatively favorable. Thus, incorporating aspects of the anomalous TC vortex structure may aid existing statistical TC intensity prediction models.