A statistical investigation of the variation of total ozone amount as measured by the deviation of ten-day running means to find out the influence of intense solar flares or of an intense solar corpuscular bombardment was carried out, each concerning with varying latitudinal belt. The method of calculation was the key day test or the superposed epoch method to see the variation from three-days before through ten-days after the key day. The ozone data used were those of twenty-two stations of the world namely, of Japan (5), America (5), Australia (3), Italy (3), Iceland (1), India (4) and of Switzerland (1), covering the period from January 1958 to March 1959. Key days were defined in two ways, namely the day when geomagnetic Ap index was greater than 45 and the day when the flare index, as defined by this author basing on the world wide flare data, was greater than 1500. According to the results of calculation ozone increased on the third or forth day after the flare key day except in the zone of lowest latitude. The statistical test did not show in general a property so much remarkable as to exceed the 5% confidence limit except the increase of ozone on the third day in the zone of 20∼30° latitude. In the case of geomagnetic key day generally none of the variation exceeded the 5% confidence limit except the case of 8 th day at the highest latitudinal belt. Also at the same zone it showed the tendency to decrease on the third day which is to be noted when compared with the increase of temperature at 100mb in the auroral zone that was reported by Ward.
Lettau's diffusion equation for surface layer with shear is modified by adding one more term and the corrected new equation is applied to a well known diffusion problem. Its solutions for large and small dispersion times are given with an experimental verification.
The error of wind speed and wind direction which are determined from the rawinsonde observations are estimated in the section 2. The probable error of wind speed is about 2m/sec at 500mb and 5m/sec at 300mb level. When the average magnitude and mean value of each term of the equation of motion are calculated by using the aerological observations, what is the accuracy of the results? This problem is discussed in the section 3. Especially, accuracy of the estimated frictional force in the free atmosphere is considered. The mean frictional force for sufficiently many cases may be statistically significant. In the section 4, the error of the vertical velocity which is derived from the windsaloft data is estimated. And a correction formula to reduce such an error is presented.
In order to obtain more information about the structure of the surface boundary layerin the atmosphere, especially in thermally stratified conditions, analyses of the data obtained at O'Neill, Nebraska are conducted. From the viewpoint of the similarity theory developed by the Russian meteorologists, the fundamental coefficients such as (∂U/∂logz)z=0, (∂T/∂logz)z=0, roughness parameter and others are determined graphically, considering that the profiles of wind speed and temperature are subject to the "log+linear" law. Then the fundamental parameters such as the friction velocity, the friction temperature, and the stability are calculated. The physical quantities derived from the fundamental coefficients, such as the shearing stress and the turbulent heat flux, are compared with the observed, values. Then the relationship between the structure of the turbulence and the stability is studied. Results are as follows: (1) The similarity theory holds. If the "constant" of the linear term is assumed to change with the stability, the "log+ linear" law is considered to hold over a wide range of stability. The calculated values of the shearing stress and the turbulent heat flux are fairly agreeable. (2) Every physical quantity is considered to be arranged only by the fundamental parameters (or the fundamental coefficients, that is, the profiles of wind speed and temperature). For instance, the power index in the "power" law, the standard deviation of wind speed divided by the friction velocity, the standard deviation of azimuthal angle and the relation of the nondimensional spectral density to the nondimensional frequency can be adjusted by the stability which is represented by the height divided by the stability length.