Evaluating Different Methods for Determining the Vertical Structure of Tropical Cyclones in Mesoscale Models

An overview and analysis of techniques used to determine the center of tropical cyclones at any vertical level in mesoscale models is presented. A variety of center-finding techniques, such as height minimums, PV centroids, and HRD's simplex routine, are explored and tested on a sample dataset of hurricane forecasts from three operational models: HWRF, GFDL, and COAMPS-TC. Eleven distinct centers, spanning the three classes of center-finding methods, plus one average center are used to calculate spread and tilt and to illustrate sensitivity of winds decomposed into radial and tangential components. It is shown that the spread of the centers is fairly consistent up to 600 mb but begins to diverge at higher levels. It is also shown that spread at higher levels is very closely related to intensity, a result which spans all three of the models. Finally, as it pertains to tilt, this talk will demonstrate how center-finding methods diverge not only in magnitude but also in direction in the vertical. A statistical analysis is performed on whether or not the differences between tilt magnitudes are significant. It is shown that for this dataset, mass-based centers give the smallest tilts at a significant level, while the tilt of PV centroids is dependent upon the weighting area.