The Predictability and Dynamics of the Overland Reintensification of Tropical Storm Erin (2007)

During the early morning hours of 19 August 2007, three days after landfall, the remnant circulation associated with Tropical Storm Erin dramatically reintensified over west-central Oklahoma. Previous studies in the peer-reviewed literature have documented the structure and environment of the reintensifying tropical cyclone; its contributions to a predecessor rainfall event across the Upper Midwest of the United States; and the contributing role of abnormally moist soil moisture conditions on synoptic-to-seasonal time scales to the reintensification of the tropical cyclone. However, the predictability of and influence of mid-latitude forcing upon Erin's observed overland reintensification remain uncertain. To that end, we utilize a thirty-member ensemble of 4 km WRF-ARW v3.4.1 numerical model simulations to investigate the predictability and dynamics of Erin's overland reintensification. The Ensemble Kalman filter implementation provided by the WRF-DART data assimilation system is utilized to provide perturbed initial and lateral boundary conditions for these simulations. The ensemble is initialized at 1800 UTC 17 August 2007, while numerical simulations begin at 0000 UTC 18 August 2007 and extend forward until 1800 UTC 19 August 2007. Physical process parameterizations are selected as appropriate for the modelling of tropical cyclones. The ensemble forecast provides a diverse range of solutions for Erin's simulated intensity during the reintensification period (Fig. 1), with minimum sea level pressure varying linearly (r = 0.974) from 997 hPa to 1005 hPa at 1200 UTC 19 August 2007. Ensemble-based statistical correlation methods are utilized to examine the sensitivity of Erin's overland reintensification to multiple meso-alpha- to synoptic-scale kinematic and thermodynamic parameters. Preliminary findings suggest that the peak intensity of the simulated Erin vortex during the reintensification period is sensitive to its simulated track; the simulated evolution of the mid-latitude synoptic-scale longwave pattern; and, to a lesser extent, the simulated structure of the nocturnal low-level jet to the south of the simulated Erin vortex. Storm-following composites for selected atmospheric fields are derived from the simulations containing the five strongest and five weakest simulated Erin vortices to further examine this sensitivity and illuminate the practical and intrinsic predictability of Erin's overland reintensification. Research findings are discussed during the presentation. Implications of these findings upon the predictability and dynamics of tropical cyclogenesis are drawn, thereby providing valuable insight into whether successful numerical model forecasts of tropical cyclogenesis events are truly skillful or are merely serendipitous in nature. Fig. 1 Caption: (a) Simulated track from each of the thirty members of the ensemble forecast of Erin's overland reintensification, as shaded by minimum sea level pressure (hPa; legend at right), valid between 0000 UTC 18 August 2007 and 1800 UTC 19 August 2007. For reference, the NHC "best track" of Erin during this time period is depicted by the solid black line. (b) Simulated minimum sea level pressure traces (hPa; grey lines) from each of the thirty members of the ensemble forecast of Erin's overland reintensification, valid between 0000 UTC 18 August 2007 and 1800 UTC 19 August 2007. For reference, the ensemble mean minimum sea level pressure trace is depicted by the solid black line. Black (red) numbers in the top (bottom) panel correspond to the five strongest (members 2, 8, 17, 18, and 22) and five weakest (members 6, 14, 15, 20, and 29) simulated Erin vortices, as assessed utilizing simulated minimum sea level pressure forecasts valid between 1200-1800 UTC 19 August 2007.