Concentric eyewall simulated in a fully compressible, nonhydrostatic, multiply nested, movable mesh tropical cyclone model (TCM4)

The occurrence of concentric eyewalls is a primary mechanism that can cause rapid tropical cyclone (TC) intensity change. Although this phenomenon has been well documented for more than two decades from observations, there have been no complete theories that can adequately explain why concentric eyewalls occur in some TCs but not in others. It is still difficult for either numerical weather prediction models or research high-resolution models to simulate such a phenomenon. This may be partially responsible for the low predictive skill of TC intensity change by numerical weather prediction models. In this study, a newly developed quadruply-nested, fully-compressible, high-resolution, nonhydrostatic model (TCM4) is used to simulate the concentric eyewalls and its evolution into the annular structure and to investigate the mechanisms responsible for the development of concentric eyewalls in the simulated tropical cyclone. The new TC model (TCM4) is an extension of the previously developed hydrostatic model TCM3 with the replacement of the hydrostatic dynamical core by a fully-compressible, nonhydrostatic dynamical core. TCM4 shares the state-of-the-art model physics, the two-way interactive multiple nesting, and automatic mesh movement with its hydrostatic counterpart TCM3. A major feature of TCM4 is its capability of simulating the TC inner core structure at very high resolutions compared to the hydrostatic version TCM3. An efficient forward-in-time, explicit splitting scheme is developed for model integration, which is a combination of a forward-backward scheme for integration of acoustic and gravity modes and a third-order upwind scheme for the three-dimensional advection terms. For an initialy axisymmetric vortex embedded in a quiescent environment on a beta-plane, the model successfully simulates the development of concentric eyewall structure and the subsequent eyewall replacement and the associated intensity change. It is found that the beta-effect modifies the radial PV structure of the model hurricanes and produces a constant asymmetric forcing for generating the convective spiral rainbands, which are axisymmetrized to form a quasi-annular first stratiform and later convective rainband–the second eyewall. The timing of formation of the second eyewall is quite stochastic and sensitive to many aspects of the model parameters. As in real hurricanes, once the second eyewall forms, the inner core convection starts weakening accompanied by a dropping of the maximum surface winds with a small increase in the minimum central surface pressure. When the inner eyewall weakens and replaced by the second eyewall, a large-eyewall may remain for several days, showing clearly the so-called annular structure. These are all consistent with the typical concentric eyewall phenomenon observed in many real hurricanes. A series of sensitivity experiments are performed to elucidate the key model parameters to the realistic simulation of hurricane concentric eyewalls.