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

This study describes ensembles of idealized tropical cyclone (TC) simulations conducted within a new modeling framework that allows for time-varying environments. These simulations are used to examine the predictability implications of exposing a TC to environmental wind profiles with different structures and magnitudes of vertical wind shear. In moderate shear magnitudes, the TCs exposed to more deeply distributed shear are weaker and exhibit a more uncertain intensity evolution than TCs exposed to shallow-upper level shear. The weaker mean intensity and higher uncertainty are found to be associated with a bifurcation in the vortex tilt response. Structural predictability time scales, computed as the time it takes for errors in the amplitude and phase of azimuthal asymmetries of radar reflectivity to saturate, are generally between 2 and 3 days for low-wave-number asymmetries. These asymmetries are slightly more predictable in deeply distributed shear and in the rainband region due to the larger vortex tilt magnitudes of many ensemble members providing more robust forcing for a quasi-stationary rainband feature. Once the magnitude of vertical wind shear increases to universally destructive levels at the end of the simulations, intensity errors begin to decrease and low-wave-number structural errors desaturate as the quasi-stationary rainband becomes a consistent feature in members of both ensembles. The most significant error desaturation occurs in TCs previously exposed to shallow-upper level shear. These results suggest that the intensity and precipitation structure are more difficult to forecast when a TC is exposed to moderate and deeply distributed vertical wind shear.