Lagrangian mean motion induced by a growing baroclinic wave is discussed, based onthe solution of Eady type problem of baroclinic instability including non-geostrophic effect. It is shown that to the leading order of Rossby number, the Lagrangian mean meridional motion is convergent toward the center of the channel. This means that air particles are mixed horizontally as a consequence of the instability. It is also shown that to the second order, air particles move downward near the northern wall and upward near the southern wall, while in the central region they move southward in the upper layer and northward in the lower, except for weak reverse flows near the top and the bottom. This Lagrangian mean picture is completely different from the usual Eulerian mean picture, and agrees qualitatively well with the result of Kida's (1977) numerical experiment as far as the behaviors of tropospheric particles are concerned, and with the result by Riehl and Fultz (1957) obtained in a rotating annulus experiment as far as the distribution of Lagrangian mean vertical flow is concerned. The result that the Lagrangian mean velocity field is convergent (divergent) even under Boussinesq assumption (cf. Andrews and McIntyre, 1978) is attributed mainly to horizontal mixing term and partly to a term of transverse-gradient transport. Eliminating the horizontal mixing term from the latitudinal component of Stokes drift and also a term of transversegradient transport from the vertical component, we can obtain the solenoidal part of Lagrangian mean meridional velocity field. This residual circulation is somewhat similar to the Eulerian mean meridional circulation, and it may be equivalent to the Lagrangian mean meridional circulation induced by a dissipating planetary wave (of. Matsuno and Nakamura, 1978). It is shown that only a part of the second order field mentioned above can be responsible to the change in Lagrangian mean zonal flow. As a result, the direction of the mean zonal flow acceleration is reverse to that in the Eulerian mean problem. Finally, we estimate the so-called eddy diffusivity, to obtain that KH=9.6×109cm2/sec and KV=8.1×103cm2/sec under the assumed condition of baroclinic wave which is chosen as a typical cyclone. It is further pointed out that latitudinal buoyancy (heat) flux consists of down-gradient transport (or particle mixing) term and transverse-gradient transport term, and that the latter is about 20% of the former in magnitude.
Resonance of lee waves produced by two barriers in a two-dimensional channel is studied with special regard to relationship between the amplitude and the distance between the two barriers. This study deals with a two-dimensional, steady flow in a stably stratified, incompressible, and inviscid fluid. The governing equation is solved analytically under the approximate boundary conditions, and the results are compared with those solved numerically by finite-difference method under the accurate boundary conditions. When the distance between the two barriers increases, flow features periodically vary with the period which is equal to the characteristic wave length λ1 of the lee wave. The maximum amplitudes of lee waves are obtained when L=L0+nλ1 (n=0, 1, 2, ... ), where L0 is a constant which depends on the Froude number.
The diurnal variation of wind at nine levels between 1000 and 100mb over eight Japanese stations is obtained from the balloon data during the period of 1956 and 1958. The result of computation indicates that the structure and behavior of the diurnal tidal motion in the troposphere and lower stratosphere are strongly affected by surface topography. This fact is also confirmed by a comparison study of vertical distributions of the diurnal meridional wind oscillation at a coastal, an inland and a mid-ocean stations in middle latitudes. The dominance of vertically propagating modes in the lower atmosphere over the coastal station, and somewhat large magnitude of trapped modes originated in theplanetary boundary layer over the inland station are found. Difference between the observation and theory is also discussed.
By the use of data at three stations close to and near the equator over the Indian Ocean area, the intensity and the zonal propagation of equatorial waves are examined by spectral analysis for the four years from November 1968 to October 1972. Spectral estimates are computed for every three months, and two types of wave disturbances are detected throughout most of the four years though the intensity of their amplitudes fluctuates: (1) Eastward propagating wave disturbances of the zonal wind component and temperature with period of about 10-20 days carrying westerly momentum upward, and (2) Westward propagating wave disturbances of the meridional wind component with period of about 4-5 days. They appear to be Kelvin waves and mixed Rossby-gravity waves, respectively. Four-year averaged spectra are computed, and a climatological description is attempted for the basic situation of these two types of equatorial waves. Biennial and annual components of wave intensity are computed by simple Fourier transforms of time series of three-monthly spectral estimates. The phase relationship of biennial components to the quasi-biennial oscillation is roughly consistent with previous studies. Kelvin wave intensity near tropopause shows a marked annual cycle with the phase of maxima in April-July, nearly in phase with the westerly acceleration in the layer 70-100 mb and with the upper tropospheric easterly acceleration. Other types of equatorial waves have not yet been detected, though it is not conclusive if the disturbances of the zonal wind component above 30 mb are Kelvin waves.
Precipitation processes and the structure of precipitating clouds in cold polar continental air mass transformed over the warm sea surface were studied on the basis of radar observations and ice nuclei measurements made around the Southwest Islands, specially over the South Okinawa region, during AMTEX periods in 1974 and 1975. It is concluded that only lower cumulus clouds were not able to produce the rainfall detected on the ground in the situation of cold air outbreak, except in relation to synoptic disturbances such as East China Sea cyclone and fronts. The rainfall which was observed with the amount more than 0.5mm per 6 hours over the South Okinawa region in the situation of cold air outbreak resulted from the coupling of lower cumulus clouds and middle-level clouds which are inferred to have been formed in association with water vapor and ice nuclei transported from the South China Continent by westerly wind.
Lateral space spectrum is obtained from a simple model based on the experimental results in high wind conditions. The logarithmic space spectrum has a peak at wave number of about 0.025m-1, which corresponds to 250m in terms of wave length, at the height of 41.5m above the ground. This wave number is approximately 3.2 times larger than peak wave number of longitudinal space spectrum. The effects of averaging time and sampling duration on lateral space correlation and spectrum are estimated from the above model. Sampling durations more than 300 seconds have effects of little significance, but averaging time has appreciable effects. Peak wave number of space spectrum varies a little with increasing averaging time when it becomes more than one second. Lateral space spectrum obtained from the above model is expressed by a simple algebraic expression. Applying this expression to the modified Davenport's theory, gust factors of fluctuations of wind velocities averaged over lateral distance are obtained. Normarized cospectrum of vertical velocities between two points with lateral separations is obtained. Using this result, lateral space spectrum of vertical velocity is obtained with the similar method to that of longitudinal velocity though this is very rough estimate due to insufficient number of data. The peak wave number of the logarithmic spectrum is nearly equal to 0.105m-1.
A method has been developed to estimate the sea-surface temperature from satellite measured infrared radiation in the atmospheric window region, taking into account a priori knowledge of statistical character of surface temperature, atmospheric temperature and atmospheric moisture. It has been shown that the information contents about the sea-surface temperature, atmospheric temperature and moisture, involved in the observed radiation in the atmospheric window region, greatly vary with the amount of water vapor in the atmosphere. In lower latitude, or in midlatitude in summer, the satellite observed radiation in the atmospheric window region contains very little information about the sea-surface temperature, and it rather contains much more information about the average temperature or water vapor amount in the atmosphere.
The atmospheric electric field was measured with a field mill at Sanriku during October 1975-October 1976. Using data thus obtained of 150 undisturbed days during the period, the annually averaged diurnal variation of electric field was deduced. It showed the UT variation with a small local-time effect. The phase of the local-time variation is shifted behind by about 2hrs from the normal local-time variation. The main sources of condensation nuclei which cause the local-time variation at Sanriku are deduced to be a train of large cities located inland about 50-90km southwest to northwest of Sanriku.
Inconsistency in writing a set of partial differential equations in finite difference form results in error in the numerical solution. An example is shown in which horizontal wind components are variationally adjusted to remove the surface pressure tendency. Three Euler-Lagrange equations are solved simultaneously using two different finite difference forms, one of which is consistent and the other inconsistent. Derived quantities (divergence and ω-velocity) are examined using both consistently and inconsistently adjusted horizontal wind components