Thermodynamical treatment of the atmosphere assumes adiabatic. change and quasi-static process of the atmosphere, and this means to assume that the effects of radiation and turbulence are negligible. _??_owever these effects, especially the latter, may not be neglected in the case of the convection of the air_??_. Applying the aerodynamical theory of the jet from circular nozzle, the author studied the effect of turbulence on the distribution of the wind velocity and the temperature in the ascending current of the air. After complicated calculations, following results were obtained. The diameter of the ascending air increases linearly upward and the stream line of the air is as shown in Fig. 2. Fig. 3 shows the distribution of the upward wind velocity and the temperature at any horizontal section. The lapse rate of the temperature is largest on the central axis of the current and is larger than the dry adiabatic lapse rate by 0-10%.
The total reflection of a mat surface such as of snow, white wall and soil, consists of two parts. The one is the reflection from the surface and the other the reflection due to the multiple reflection inside the medium. This paper treats the problem of the reflection merely from the surface itself. We denote the angle between the normal of any facet, which is a component of the_??_mat surface, and the normal of the mean surface, by θ. As the distribution function W(θ) of the directions of the facets, we take the form, following the line of the discussion by L. S. Ornstein.(2) However, contrary to Ornstein, who assumed that every facet reflects the light totally like a mirror, the present author hag assumed that reflectivity, _??_, obeys Fresnel's formula. After some calculation, we have obtained that R_??_. the partalbedo due merely to the surface itself, is given by when the incident light is of parallel nature and has the incident angle In the above equation, _??_, i and θ are given by the equations (1), (6) and (7) in the original Japanese paper by the author, respectively. When the incident light is of diffused nature, we make use of the relation between M_??_ and M_??_, where means the intensity of reflection of _??_ -direction when the incident light is diffuse, and M_??_means the total amount of reflection when the incident light is of parallel nature and of the incident angle _??_ . This equation was formerly obtained by M. A. Boutaric, (4) and has been re-obta-ined in much simpler way by this author in the present article. Making use of Boutaric's relation, it has been proved that, for the incident light of diffused nature, the albedo of the reflection due merely to the surface itself, is given by the equation.
As a theory _??_iscussing the structure of the lower atmosphere there is J. Bjerknes' Oell Theory. According to this theory the subtropical belt of high pressure is divided into separate anticyclones, or cells, and the longest axis of a subtropical cell is of the order of 4000 to 9000km. But, in practice the anticyclones which appear in our synoptic charts covering the region in which the growth and invasion of the typhoons take place are sometimes by far the smaller than the Bjerknes' cell and their dimensions are of the order of 1000km or less. To explain this fact, we propose a theory alike to the mosaic structure theory due to Dr. S. Fujiwhara. Its principal conception is that the lower atmosphere is divided into cellular anticyclones of the order mentioned above and each cell is comparatively stationary. If we admit this mosaic theory, the growth and movement of typhoons are much more clearly explained. For intance, the growth of typhoons which originate in the vicinity of latitude 22°N are difficult to be explained merely by the equatorial front, for in such cases the equatorial front is clearly traced near 10°N. The typhoons in such cases are originated on the front which runs nearly parallel to the latitude of 22°N (tentatively named tropical front) and the existence of this front can rationally be shown merely by admitting the mosaic structure theory.
The author studied statistically the relation between moving cyclones and the weather at Ootomarl, Sikka, Maoka and Anbetsu in Southern Sakhalin, and also monthly frequency classified according to the depth cyclones passing through this district and the neighbourhood. It was found that the weather conditions are determined by season, quadrantal position of the centre, distance from the centre and pressure difference between one place and the centre of the cyclone. Thus it became clear that there are very striking local cheracteristics on the weather condition caused by the passage of cyclones in southern Sakhalin. (especially at the eastern and western coasts of this island), and some practical and useful conclusions for the weather forecast were obtained.