This paper describes results from numerical experiments which have been made toward a better understanding
of tropical cyclone formation. This study uses a nonhydrostatic version of the author’s mesoscale-convection-resolving
model that was developed in the 1980s to improve paramerization schemes of moist convection. In this study the
horizontal grid size is taken to be 20 km in an area of 6,000 km x 3,000 km, and a non-uniform coarse grid is used in two
areas to its north and south.
Results from two numerical experiments are presented; one (case 1) without any environmental flow, and the other (case
2) with an easterly flow without low-level vertical shear. Three circular buoyancy perturbations are placed in the west-east
direction at the initial time. Convection is initiated in the imposed latently unstable (positive CAPE) area. In both cases, a
vortex with a pressure low is formed, and two band-shaped convective systems are formed to the north and the south of
the vortex center. The vortex and two convective systems are oriented in the westsouthwest – eastnortheast direction, and
their horizontal scales are nearly 2,000 km.
In case 1, the band-shaped convective system on the southern side is stronger, and winds are stronger just to its south. In
contrast, in case 2, the northern convective system is stronger, and winds are stronger just to its north. Therefore, the
distributions of the equivalent potential temperature in the boundary layer and latent instability (positive buoyancy of the
rising air) are also quite different between cases 1 and 2. The TC formation processes in these different cases are
discussed, with an emphasis on the importance of examining the time change of latent instability field.