In section 2, we compute t h e height changes depending only upon the horizontal advection of absolute vorticity at each isobaric level, such as 300-mb,500-mb,700-mb, and 850-mb. The location of negative or positive height change centers at each level varies from level to level, and the so-called non-divergence level differs with individual typhoons. Referring to this fact, it may he said that the barotropic model is not always a sufficient one for the forecasting of typhoon movement.
In section 3, the simplest method for taking into account the baroclinic effect is treated. Assuming the vertical distribution for ω to be parabolic, ω, and also ∂ω/∂
p, are easily obtained by using the data at two selected isobaric levels. Examining the height change due to ∂ω/∂
p at various isobaric levels, it becomes clear that the parabolic assumption for ω is not always adequate for the forecasting of typhoon.
In section 4, the more complete model for baroclinic case is treated. Solving the differential equation of vertical velocity by taking up as many isobaric levels as possible, such as 300 mb,500 mb,700mb,850 mb and 1000 mb, we analyse the detailed vertical distribution of ω. The distribution of ω is somewhat complicated, especially in the lower atmosphere. It may be guessed for certain that the baroclinic factor ∂ω/∂
p effectively contributes to the movement of typhoon.
In section 5, the height changes for next 12 hours is de s cribed. The patterns of these changes involve considerable irregularities whose cause is not computational instability but truncation error or error of physical nature.
In section 6, to a void these errors, a tentative method for smoothing the initial pattern is proposed. This method is a fairly satisfactory one, although it would not be the best.
View full abstract