In the succeeding chapters we shall endeavour to develop, on the principles of air mass analysis, the dynamical climatology of Japan and show how these principles may be applied to the problems of weather chart analysis and weather forecasting. The physical analysis of the weather charts, naturally, forms the basis of rational forecasting. This theory necessary presupposes the existence of well-defined air masses, distinguishable by certain dynamically significant conservative or nearly conservative properties. Otherwise little meaning could be attached to the fronts. For the Japan proper, the trajectories of the air masses involved are very long. The Japan proper lies to the east of the Asiatic continent surrounded by the seas and ocean and stands under the decided influence of the continent to the west. Although the source region of continental air masses is extensive enough to make possible very large scale of currents of this origin, the air masses come to us usually after some days spent in oceanic region, a fact which helps to explain why most of their primal characteristics are modified so quickly. Thus little meaning could be attached to the fronts and it is doubted by those who have considered the situation that the same advantageous results as in western Europe and in North America can be obtained from air mass analysis over areas where predominatingly continental climate obtains. It is the reason why the majority of the daily forcasting services throughout Japan is using empirical method based primarily on migratory barometric pressure system with some modification to the Norwegian methods. In recent years, however, the so-called Norwegian system of air mass and frontal analysis has been employed with considerable success in this country. The threedimensional nature of the analysis permits of ready adaptation of the requirements of weather prediction. Many of the conclusions reached herein are not new to meteorology and climatology, but heretofore have lacked confirmation by observations in the upper air. Because these data were obtained in the period where there were fewer aerological ascents for meteorological purposes than there are at present a sufficient number of representative soundings for the fairly well-defined and fast moving air masses can not been obtained. While the conclusions reached may have been influenced by this scanty material, it is believed that the results are generally applicable. It will be noticed at once that the meteorologist using air mass and frontal method must, from experience, be familiar with both the source properties of the air masses and the typical transformation of each source type to the transitional form. The determination and explanation of these source properties and typical transformations constitute the main subject of discussion in the present paper.
In the present paper, the following results are reported. At first, the actual computation, having due reference to different boundary conditions, of the expressions for the circulation, gives different results. The boundary conditions were assumed by A. Oberbeck and others that at the upper limit of the atmosphere slipping occurs without friction but the motion at the earth's surface is everywhere zero. The last condition, however, is doubtful. The other extreme conditions that both at the upper limit of the atmosphere and at the earth's surface slipping occurs without friction are assumed and the solutions are obtained. The results obtained under these assumptions agree with the atmospheric phenomena in their principal points, but do not agree in several points. The probable boundary conditions should be expected to be some intermediate ones between these two extreme cases. Next, the effect of the centrifugal force due to the earth's rotation is discussed and concluded that in order to introduce the effect of the centrifugal force one must consider the earth as a ellipsoid of revolution instead of a sphere. Next, the secondary circulations as monsoons are treated by introducing further terms that depend on the geographical longitude, also. Lastly, it is concluded that the numerical coefficient of eddy viscosity is larger than 107 C. G. S. of the order of magnitude for the general circulation of the atmosphere. Beside the ordinary circulation, the circulation over polar region being northward below but southward above: descending in the temperate zones but ascending in the arctic region in the northern hemisphere is derived theretically.
In the previous paper on the same problem, the present author has treated an example of Tunami or destructive sea waves in the bay of Oosaka excited by the Muroto typhoon on the 21st, September 1934. It should be expected that the height of Tunami reaches a maximum value when the velocity of typhoon is almost equal to the value of √gh, where g is the acceleration due to gravity, h the mean depth of the bay, to be derived from the period of the first order of the harmonics of the oscillation. To visualize this point, the author tried to solve the mathematical expressions given in the previous paper, by which we can characterise the height of Tunami excited in an arbitrary rectangular bay by any travelling disturbance, for the bay of Oosaka and for each typhoon with the velocity of 30, 50, 60, 70, 90, and 150km/hour respectively. The result derived from the treatment will be easily understood from the Figure 1; the maximum height of tunami varies according to the velocity of the typhoon and it has a maximum point for the typhoon having the velocity of about 60km/hour, which is almost the same with that of √gh calculated for the bay of Oosaka and the maximum height occurs a little later after the arrival of the typhoon centre. Some additional characteristics will easily by seen from the Figures 1 and 2.
The mathematical treatment of the waves in viscoelastic medium has been given by Dr. K. Sezawa and some others when they damp with time. The present author treats of the several cases when they damp with distance.