Initialization and Simulation of a Landfalling Hurricane Using a Variational Bogus Data Assimilation Scheme

摘要

The Bogus Data Assimilation (BDA) scheme designed to specify initial structures of tropical cyclone was tested further on the simulation of a landfalling hurricane - - Hurricane Fran (1996). We studied the sensitivity of the simulated hurricane track and intensity to the specified radius of maximum wind of the bogus vortex, the resolution of the BDA assimilation model, and the bogus variables specified in the BDA. In addition, we compared the simulated hurricane structures with the available observations, including the surface wind analysis and the radar reflectivity captured on-shore during Fran's landfall. The sensitivity study of the BDA scheme showed that the simulations of the hurricane track and intensity were sensitive to the size of the specified bogus vortex. Hurricanes with a larger radius of maximum sea-level pressure gradient are prone to a more westward propagation. Results of the hurricane initial structures and prediction were also sensitive to the bogus variables specified in the BDA. Fitting the model to the bogused pressure data reproduced the hurricane structure of all model variables more efficiently than when fitting it to bogused wind data. Examining the initial conditions produced by the BDA, we found that the generation and intensity of hurricane warm-core structure in the model initial state was a key factor for the hurricane intensity prediction. Initialized with the initial conditions obtained by the BDA scheme, the model successfully simulated Hurricane Fran's landfall, the intensity change, and some inner-core structures. Verified against the surface wind analysis, the model reproduced the distribution of the maximum wind streaks reasonably well. The model-predicted reflectivity field during the landfall of Hurricane Fran resembled the observed radar reflectivity image on shore.