Abstract
A variety of tropical-cyclone (TC) center-finding methods aggregated from previous works of mesoscale modeling and operational analysis are compared. The previous methods used can be divided into three classes: local extreme, weighted grid point, and minimization of azimuthal variance. To analyze these methods, four representative separate TC forecasts from three operational models—the Coupled Ocean–Atmosphere Mesoscale Prediction System Tropical Cyclone version, a Geophysical Fluid Dynamics Laboratory model, and the Hurricane Weather Research and Forecasting Model—are examined. It is found that for this dataset the spread of the derived TC centers is fairly small between 1000 and 600 hPa but begins to increase rapidly at higher levels. All models exhibit increased center spread at upper levels when the TCs’ strengths fall below approximately hurricane strength. On a given pressure level, tangential wind differences calculated from different centers are generally small and localized, whereas radial wind differences are often much larger in both space and relative magnitude. Center-finding techniques that use mass fields to calculate centers exhibit the smallest vertical tilts for hurricane-strength TCs. Conversely, potential vorticity centroids with large weighting areas produce the largest tilts. Given the potential sensitivity of center determination and implied tilt for various other measures of TC structure (radius of maximum winds), these results may have large repercussions on both past and future analyses.
