Examining Trends in Satellite-Detected Overshooting Tops as a Potential Predictor of Tropical Cyclone Rapid Intensification

A geostationary satellite-derived cloud product based on an overshooting top (OT) detection algorithm is described for applications over tropical oceans. Tropical OTs are identified using a modified version of a mid-latitude OT detection algorithm developed for severe weather applications. This algorithm is applied to identify OT activity associated with Atlantic tropical cyclones (TCs). The detected OTs can serve as a proxy for “hot towers”, which represent intense convection linked to TC rapid intensification (RI).

This study will describe the OT algorithm development and examine its potential as a predictor both on its own and in multi-parameter RI forecast schemes. RI forecast skill potential is explored by examining thresholds of OT activity and trends within prescribed radii of TC centers. An independent test on Atlantic TCs in 2006-2007 reveals an empirically-based OT scheme has potential as a predictor for RI beginning in the subsequent 24 hours. While not being able to skillfully predict the ultimate length of the RI period, this scheme shows a promising probability of detection and false alarm ratio. Further testing also suggests the OT algorithm is skillful when predicting if RI will occur and complete within the subsequent 24 hours, especially for RI maximum wind thresholds of 25-kt/24hr and 30-kt/24hr.

However, as expected, the stand-alone OT-based RI scheme is comparatively less accurate than existing objective multi-parameter RI prediction methods. Therefore, a preliminary experiment adding OT-based predictors to an objective logistic regression-based scheme is shown to slightly improve the forecast skill of RI.