The Structural Predictability of Idealized Tropical Cyclones in Environments with Time-Varying Vertical Wind Shear

This study concerns the structural predictability implications of exposing a tropical cyclone (TC) to different magnitudes and distributions of vertical wind shear in ensembles of idealized simulations with time-varying environments. The measures of TC structure include the amplitude and phase of azimuthal asymmetries in simulated radar reflectivity averaged within two radial bands. We assess structural predictability by comparing errors, computed as root-mean-square differences in the structural measures between the perturbed ensemble members and the unperturbed simulation, to some basic statistics of TC structure within the ensembles. In moderate shear environments, the precipitation structure remains predictable until about 12 hours for the smallest-scale asymmetries considered, and until about three days for the wavenumber-one asymmetry and the azimuthal mean, with slightly longer predictability in the rainband region than in the eyewall. Predictability timescales exhibit a weak relationship to the shape of the wind profile different from what we expected. Both amplitude and phase errors saturate earlier when TCs are exposed to the less destructive wind profiles with shallow upper-level shear. However, as the shear magnitude increases, the TCs in shallow upper-level shear exhibit a more pronounced recovery in structural predictability. The phase of precipitation asymmetries has similar predictability timescales as the amplitude, except phase errors saturate more rapidly with increasing azimuthal wavenumber and the phase of the wavenumber-one asymmetry exhibits a more pronounced error desaturation in strong shear. The evolution of precipitation structural errors within these ensembles confirms some of the inherent difficulties in forecasting TCs in moderate shear environments.