Initial values of horizontal and vertical wind components, temperature, pressure, and moisture are objectively determined for use in a boundary layer forecast model. The objective analysis is accomplished by use of Sasaki's (1969 a, b) numerical variation method. This technique incor-porates the governing dynamical and thermodynamical equations, as well as observations, into the initialization process. The solution to four coupled partial differential equations with asso-ciated boundary conditions completely determines the initial map. Richardson's relaxation method is used to solve the system of equations. The initialization and prediction is applied to severe weather occurrence in the Midwest on 10 June 1968. A 1.5km thick layer bounded by the earth's surface and encompassing a horizon-tal area of approximately 2000×2000km2 is used. The horizontal grid spacing is 190km, half of the NMC (National Meteorological Center) grid, and the vertical interval is 200 m. Two sources of data are used in conjunction with the governing prognostic equations to determine the initial fields: 1) wind and specific humidity observations from the radiosonde network, 2) pressure and temperature fields from the NMC analysis. The pressure and tem-perature data could also be extracted from the radiosonde records to obtain a more consistent set of data. The hybrid collection of information, however, helps reveal the versatility of the variational method. The degree of confidence in each data set is controlled by appropriate choice of weights in the variational formulation. The areal distribution of a severe storm index, wq, the product of the vertical velocity and specific humidity, at the mid-level of this model is discussed. The initial distribution and the 3hr and 6hr forecasts of the index are compared with the surface observations of severe weather.
Laboratory experiments have been made to investigate the dependence of finite-amplitude Rayleigh convection cells on the initial conditions of the system. Controlled perturbations with prescribed width-depth ratios were introduced into the fluid (silicon oil) confined in a box which had a small breadth. Subsequent evolution and final steady flows at the Rayleigh number of approximately 45 Rc were observed, where Rc is the critical Rayleigh number. It was found that "two-dimensional" convection cells showed tendencies to adjust their sizes toward the vicinities of the width-depth ratio of 2.3, provided their initial width-depth ratios were outside the range between approximately 2 and 3.
The barotropic model with the non-divergent stream function derived from wind analysis as input is used for predicting 500mb level flow patterns in the Indian region. Appropriate boundary conditions for computation of the stream function from the observed wind information are pre-scribed. Such boundary conditions are considered suitable for the purpose which maximise kinetic energy in the reconstructed wind and minimise root-mean-square-vector-error between the recon-structed wind and the observed wind. Using data of a monsoon situation in June 1966, in which a special interest is movement of a monsoon depression from the Head of the Bay of Bengal in a northerly direction, integration is performed in one-hour time step upto a period of 48-hrs. The 24-hrs. forecasts of the flow patterns appear to compare favourably with the observed flow patterns while the 48-hrs. forecasts appear to show considerable departure from the observed patterns. The r. m, s. v. e. between the forecast and observed values of the non-divergent component of the winds for the Indian region are found to be 3.5m/s for 24-hrs. forecasts, and 6.0m/s for 48/hrs. forecasts. The mean vector error between the predicted and the observed positions of the center of the deep depression is 140km for 24-hrs. forecasts, and 205km for 48-hrs. forecasts.
From considerations based on similarity it is found that the ratio of the components of the eddy diffusivity tensor of semiempirical turbulent diffusion theory can be calculated by Kxx/Kzz=(u2)2/(w2)2, Kyy/Kzz=(v2)2/(w2)2 and Kxz/Kzz=-(uw)2/(W2)2 where u, v and w represent the velocity fluctuations in the direction of the mean wind, x, in the lateral, y, and in the vertical direction, z, respectively. The result for Kyy/Kzz is in agreement with that calculated for neutral conditions by Yamamoto and Shimanuki (1964) from data obtained during projects Prairie Grass and Green Glow. The result for Kxz/Kzz has the same order of magnitude as that derivable from turbulent heat flux measurements by Zubkhovskii and Tsvang (1966).
Diffusion of particles having an equal fall velocity which are released from a cross-wind line source in diabatic atmospheres is studied. The wind profile and eddy diffusivity assumed in this investigation are those proposed by Yamamoto and Shimanuki as a revision of the KEYPS formula. The solution of the resulting diffusion equation is obtained numerically for various combinations of the source height, thermal stability and terminal velocity of the articles. In addition the deposition rate of particles on the ground is estimated as a function of the downwind distance, source height and stability. Diffusion of particles having different sizes is then estimated.
Conventional instruments are not available for measurement of air temperature at remote places, or for the spatial mean of the component for a certain direction of wind speed at a certain height above the ground over a line of some hundred meters or more. The possibility of the use of sonic means for this purpose was investigated, through the development of a sonic model device and its field test. The model device consists of a 20-W output sound transmitter, a receiver and a computing apparatus with a recorder. The field test was carried out along the path of the transmitter and receiver horns at 2.5m height above the ground with 550m baseline length for air temperature and 300m for wind speed component. The sound transit time is measured and then converted automatically into spatial means of air temperature and of wind speed component with the compensation for the effect of water vapor pressure. The products are then corrected manually for the effects of wind speed the vertical ambient air emperature gradient. For reference measurements, thermister thermometers and three-cup anemometers were placed along the baseline, whose measurements were referred to for verification of those by the sonic device. From the results of the field test, it became clear that the measurement of the mean air temperature several meters above the ground by the sonic method must be based on a more accurate knowledge of the mechanism of sound propagation and that the baseline length should be less than 550 m.
In the atmospheres of helium-argon mixture at normal pressure with varying mixing ratio of the gases, ice crystals were formed by seeding of silver iodide smoke. Crystals were grown in a state of gravity fall and fell to the bottom of a cold chamber. These crystals were examined directly without the use of replication technique by a microscope that was set under the bottom window of the chamber. It has been revealed that features of crystal growth vary with mixing ratio of helium and argon. The revealed variation of growth with respect to the mixing ratio of the gases can not be attributed to the difference in a state of fall of the crystals in the chamber because the molecular viscosity varies little with the mixing ratio of helium and argon. Instead, that can be attributed to the difference in the diffusion coefficients of water vapor and the thermal conductivities of the carrier gas, because the both parameters vary greatly with the mixing ratio of the gases.
An observation of the atmospheric electric field strength was carried out by means of a Field Mill type electrometer at Syowa Station (69°00'S, 39°35'E), Antarctica, from the beginning of Feb. 1968 to the end of Jan. 1969. As a result, the following interesting data were obtained: (i) no remarkable diurnal variations in undisturbed electric field, (ii) a positive correlation between the positive electric field and wind velocity, (iii) undisturbed electric field accompanied by southerly winds, (iv) an electric field accompanied by drifting and blowing snow, (v) a difference of the commencement time of the increase of a positive electric field and the increase of wind velocity, (vi) a serrated pattern of positive electric field and (vii) a disturbance of the electric field as influenced by the aurora.
The vertical profiles of 222Rn (radon) and 220Rn (thoron) were obtained analytically under some conditions of the diffusion coefficient with respect to altitude. The theoretical relationships between the concentration of the emanations and wind speed were compared with those obtained by field measurements, and fairly good agreements were obtained. The exhalation rates of 222Rn and 220Rn were also estimated by comparing theroretical relationships with the observed data. It was estimated as 1.2×10-16Ci/cm-sec for 222Rn and 9.2×10-16Ci/cm2-sec for 220Rn on the campus of the university.