In winter fogs, when the polluted air or liquid fog came over, the potential gradient increased positively to several times of normal value. This increase in the potential gradient was explained by the decrease of the air conductivity. Accompanying with the glaciation of those liquid fogs, the positively enhanced potential gradient decreased to the normal values or even negative values. About the decrease due to the glaciation following mechanisms are considered. Ice crystals carried preferentially the negative charge. The conductivity of the air increased by the removing submicron aerosols due to the growth of ice crystals (Facy effect). At times, it was observed that when the snowfall stopped, the potential gradient abruptly increased to positive high values. About the mechanism, it was considered that the residual positive space charge was gravitationally separated from the negative snow crystals, then the positive space charge was brought down to the surface by downdraft between snow clouds.
Various electrical properties of winter fog were measured at Asahikawa in 1968, following the previous observation as reported in Part I (Endoh, Iwabuchi and Magono 1972). From the results of measurement of the vertical distribution of the potential gradient in a fog layer, it was found that the electric potential gradient in a fog layer was greater than that above the layer. This was considered to be caused by the smaller conductivity of the air in the fog layer. When a snow cloud passed overhead during the fog, a high negative point discharge current was observed. This was explained by the negative potential gradient at the point due to the negative space charge of the snow cloud. Peculiar changes were noticed in the antenna-earth current, and they were explained by the effect of the smoke from steam locomotives, flapping of crows' wings, falling off of frost pieces from the antenna line, and the sheathing-layer-like phenomenon on the upper surface of the fog layer.
The size distributions of natural aerosols and those of radioactive aerosols have been determined simultaneously based on mobility measurements on those charged particles. The apparatus is a cylindrical aspiration chamber. In the case of radioactive aerosols, the central electrode of the condenser is essentially a gas flow counter. The counter covers the size range with approximate limits of 1×10-5 and 0.8×10-6cm in radius. The results indicate that the maximum number concentration of aerosols is located in this size region, and the distributions of radioactive aerosols are shifted toward larger size region compared with those of natural aerosols. This tendency is also marked with a number of measurements of the mean radius obtained by a diffusion battery. The size distributions of radioactive aerosols computed theoretically from those of natural aerosols are compared with the results obtained by direct measurements. The experimental results agree with the theoretical distribution as a rough first approach. But, the available data seem to indicate that the experimental results are shifted toward somewhat greater sizes than the theoretical ones.
Measurements have been made of the pressure dependence of the electric field strength Fo required to initiate corona from water drops of spherical radius R and charge Q falling between electrodes at a velocity of about 4.5m.sec-1. At a pressure p of 1000 millibars in a positive vertical field a value of Fc of 9.0±0.5kV/cm was required to produce charging of the drops, whereas for p=500 millibars the critical value of Fc was reduced to 5.5±1.0kV/cm. In the presence of a horizontal electric field the critical values of Fo at 1000 and 500 millibars were 6.3±0.3 and 6.9±0.3kV/cm respectively. Equations are developed, based principally on the work of Dawson (1969) and Richards and Dawson (1971), which explain these results and predict that the maximum fields that can exist inside the regions of thunderclouds containing large raindrops will be independent of p at around 12/R0.3e.s.u., corresponding to about 500kV/m. It is suggested that when such fields have been produced inside thundeclouds the emission of corona from disrupting drops will trigger the lightning discharge.
A method of estimating evaporation from a relatively deep lakes using the climatological data of wind velocity, air temperature and vapor pressure over lake is developed. Merit of this method is that these data are easy to infer from those at the meteorological stations in the vicinity of the lake, even if there are no direct observations over the lake. The method consists of solving the equation of heat conduction in the lake together with the heat balance equation at the water surface taking into account some results of the turbulent transfer theory, and thus obtaining evaporation and water temperature simultaneously; the surface value of the latter being deeply related to evaporation. Estimation of these values by this method is made on representative lakes in Japan and the estimated results are compared with the observed ones. In addition, some climatologicalfeatures of lake evaporation are pointed out with the help of the estimated values.
A statistical study is made for the intermediate-scale disturbances reported by Murakami (1971) from the analysis of precipitation. In this study, method of spectrum analysis is applied for the data in the Marshall Islands obtained in 1956 and 1958. It is re-examined and confirmed that these disturbances of about 3 days period propagate eastward and have a wavelength about 1, 000km. In agreement with this, it is shown that the disturbances accompany small dimensional area of wind convergence. The analysis also shows that the variation of wind velocity associated with the disturbance is prominent in the divergent part of the wind. By the method of inter-level cross-spectrum analysis, the vertical structure of the disturbances is obtained. The feature of cold core appears in the lower levels. The feature of warm core in the upper levels is clear in 1956, but not in 1958.
The properties of small-amplitude convection in the conditionally unstable stratification of the atmosphere which have been studied by Kuo (1961, 1965) are re-examined by the use of the finite difference method which implicitly introduces the condition for the continuity of the horizontal gradient of temperature at the internal boundary between an ascending region and adescending region. This condition, which was not used by Kuo, is required for the finiteness of the horizontal eddy thermal diffusivity term which is assumed to be of the Fickian type. It is shown that when the horizontal size of a convective cell as well as some physical parameters are specified, the horizontal size of an ascending area is uniquely determined. This property is markedly contrasted with the property of convection for the non-viscous case (Haque, 1952; Lilly, 1960) where the size of an ascending area is not uniquely determined. The use of the finite difference method enables us to discuss the properties of unstable convection. The dependencies of the growth rate and other properties of convection on physical parameters are studied. The properties of convection is also studied as an initial value problem. Furthermore, it is shown that this approach is very useful to obtain an eigensolution, particularly when the separation of variables with respect to the horizontal and vertical coordinates is impossible.
The problem of separation and analysis of traveling waves and fluctuations of stationary waves is discussed. Difficulties and ambiguities in interpretation of the results of this kind of analysis are pointed out, with particular reference to the method recently published in this journal by Iwashima and Yamamoto (1971).
The results of some measurements of the horizontal and vertical fluctuations of the wind as well as Pasquill's stability categories are presented. It is suggested that the definition of stability be adjusted, taking into account local geographical and climatological conditions, thus obtaining agreement between the systems of Pasquill and of Cramer.