Book Description

The dynamics and predictability of various aspects of tropical cyclone track and intensity forecasting are explored through the use of real-time convection-permitting ensemble forecasts generated by a regional-scale model that employs advanced data assimilation techniques. Airborne Doppler radar observations, as well as sounding observations gathered during NASAs Hurricane and Severe Storm Sentinel (HS3) are assimilated and the resulting sensitivity and uncertainty of divergent track and intensity forecasts for three Atlantic tropical cyclones (TCs; Hurricane Sandy (2012), Hurricane Nadine (2012), and Hurricane Edouard (2014)) are explored. Ensemble members are separated into groups according to their performance and composite analyses and ensemble sensitivity techniques are employed to diagnose the sources of greatest sensitivity and uncertainty, as well as to dynamically explain the divergent behavior observed in the forecasts.The analysis of the Hurricane Sandy (2012) ensemble reveals that the divergent track forecasts result from differences in the location of Sandy that develop over the first 48-h of the simulation as a result of variance in the strength of the environmental winds that Sandy is embedded in throughout this period. Disparities in the strength and position of an approaching mid-latitude trough yield divergence in track forecasts of Hurricane Nadine (2012); an increased interaction between the mid-latitude system and the TC steers Nadine eastward, while a reduced interaction allows the TC to be steered westward ahead of the approaching trough. In addition, the inclusion of 6-h sea surface temperature (SST) updates considerably improves Nadines intensity forecasts, highlighting the importance of accurate SST fields when simulating TCs embedded in marginally favorable environmental conditions. Finally, considerable variance in the rapid intensification (RI) onset time in the Hurricane Edouard (2014) ensemble results from small distinctions in the magnitude of deep-layer shear prior to RI, which contributes to differences in the vortex tilt magnitude, the strength and location of the inner-core convection associated with the developing vortex, and the subsequent precession process.