Recently, theoretical study of the maximum intensity, which is measured by the maximum wind velocity, of tropical cyclones (TCs) rapidly developed. In particular, the theoretical TC models having two-dimensional axisymmetric structure present the structure and intensity of real TCs very well (e.g. Emanuel's MPI theory; Emanuel, 1986). These models enable us to estimate the TC maximum potential intensity with a few environmental parameters, such as sea surface temperature and tropopause temperature in the ocean area. To support theory, a number of numerical studies of the TC maximum intensity have also been conducted. Using two-dimensional axisymmetric model, the TC maximum intensity has been predicted with complicated physical processes (e.g. turbulence and cloud physics) which are neglected in theoretical studies.

However, two-dimensional axisymmetric models do not include non-axisymmetric structure in TC. Wang (2002a,b) indicated that non-axisymmetry in an idealized TC influences the intensity and structure of TC, using a three-dimensional hydrostatic model. This suggests that a three-dimensional model is necessary to predict more accurate maximum intensity of TC. In addition, we believe that non-axisymmetric effects should be included in axisymmetric theoretical models.

The present study used a three-dimensional non-hydrostatic model named the Cloud Resolving Storm Simulator (CReSS) to estimate quantitatively how non-axisymmetry influences TC maximum intensity by analyzing the angular momentum budget in the inner core region. We analyzed the angular momentum budgets for TCs which have the different maximum intensity. The present study showed that non-axisymmetric components have negative contributions for the maximum intensity (i.e. deceleration for the maximum wind velocity) in the inner core. In addition, we showed that this deceleration due to non-axisymmetric components is about 10% of the acceleration due to axisymmetric components and the deceleration increases with the maximum wind velocity. Therefore, we infer that non-axisymmetric components should be considered for prediction by the MPI theory for intense TCs. This result suggests that the effect of non-axisymmetric components should be in cluded in theoretical models of the maximum potential intensity. Moreover, we analyzed the energetics of maintaining mechanism of non-axisymmetric components by energy budget. We showed that energy of non-axisymmetric components is supplied by conversion of potential energy of non-axisymmetric component to kinetic energy and conversion by barotropic and baroclinic processes. We can explain horizontal displacement of non-axisymmetric components as propagation of vortex Rossby wave in an axisymmetric vortex. It is important for understanding the TC maximum intensity to investigate the propagation characteristics of this wave. To show that non-axisymmetry in the inner core is propagation of vortex Rossby wave, we compared simulated waves with theoretical value of the dispersion relation of vortex Rossby wave in shallow model. We detect tangential low-frequency waves, which have a rapid propagation speed and a feature of the vortex Rossby wave in the inner core. The waves have phase speed which is comparable to the theoretical value. This result suggests that theoretical value of phase speed is identical to that of the simulated TCs in the non-hydrostatic model and in the inner core non-axisymmetric component can explain as propagation of vortex Rossby wave.