In order to study the role of surface friction in a tropical cyclone described in a numerical experiment (Yamasaki, 1977) in which the effects of cumulus convection were not parameterized but explicitly calculated, numerical experiments are performed by changing drag coefficient concerning surface friction. It is shown that surface friction plays important roles to reduce the horizontal scale of the disturbance and to produce an eye and eyewall as observed in the tropical cyclone. A more important result is that a disturbance may develop even when surface friction is not taken into account. This result indicates that Yamasaki (1977)'s tropical cyclone (referred to as AF mode) is different from tropical cyclones (TC mode) described by Ooyama (1969) and Yamasaki (1968) in the sense that TC mode becomes neutral in the absence of surface friction. It is possible to interpret AF mode as a superposition of TC mode and NF mode, where NF mode is defined as a CISK mode whose instability does not depend on surface friction. That is, when surface friction does not exist, NF mode is dominant because TC mode is neutral. Also, even when surface friction exists, TC mode is not noticeable if the vorticity of the disturbance is too weak to produce appreciable frictional convergence. When the disturbance becomes strong, TC mode with pronounced frictional convergence becomes dominant. In this case AF mode is similar to TC mode. It is also suggested that tropical cyclones simulated by Anthes et al. (1971) and Kuri-hara and Tuleya (1974) are similar to AF mode rather than TC mode.
Numerical experiments have been conducted to investigate methods of the four-dimen-sional data assimilation. After the satellite data are collected, an intermittent method of the four-dimensional data assimilation is applied, using the operational Northern Hemispheric Primitive Equation Model of the Japan Meteorological Agency. The response of the model to the inclusion of the satellite dataa is studied. The high frequency mode generated by the data insertion is well suppressed, but the low frequency mode is not completely filtered out. The impact of the satellite data seems to be acceptable in the extra tropics. However, the situation is different in the tropics; in some cases noises caused by the data insertion are not negligible in the assimilated field. The problem to suppress these noises in the tropics remains unsolved.
To analyse features of maritime cold fronts measurements of many meteorological variables were made at six levels from the NCAR Electra aircraft flying through a cold front moving southeast from Okinawa during AMTEX. The front was relatively weak and atypical in that it was double and no middle or high clouds existed above a stratocumulus shield at 2km. The history of the fronts is given and the structure described. Characteristics of several frontal air masses are determined from flight and radiosonde data. Mixing, primarily vertical by cumulus convection, between the various air masses is demonstrated. Regions of upward motion ahead of the fronts located by aircraft inertial navigation system measurements and from divergence of winds agreed well and confirmed the double structure of the front. Local areas of frontogenesis are found ahead of each front. Averaged over the entire system, weak frontolysis was occurring primarily as a consequence of the warming of air behind the fronts. Both positive and negative fluxes of heat and water vapor occurred within the frontal zone as a result of turbulent transport, direct cumulus convection, over-shotting of equilibrium levels, and subsidence of cloud remanents. Normalization of velocity spectra in the frontal zone by height times energy dissipation and frequencies by height over aircraft speed gives a universal spectral pattern indicating the dominance of generation of turbulent energy by shearing processes. Spectra observed within cold air behind the front fits the shapes and non-dimensional values characteristic of well-mixed air heated from below and capped by an inversion, indicating the dominance of heating as a source of turbulent energy.
The eruption of the Fuego in Guatemala injected a large amount of volcanic sub-stances to the lower stratosphere in mid-Oct. '74. About a month later, an extraordinary intense scattering layer was observed by ruby lidar in the lower stratosphere over Fukuoka. The peak value of the non-molecular radar cross section showed a maximum in Dec. '74. It decreased gradually after spring '75 and attained a secondary maximum in early autumn of the year. The height of the layer peak was located a few kilometers below the normal Junge layer and the half width of the layer which was initially a few kilometers increased as the peak value of the non-molecular radar cross section decreased. The gross feature of the temporal variation of the peak value observed by lidar is in contradiction to the results expected from the 2-dimensional transport models proposed up to this time and the contradiction can not be explained if we consider only the case where the observed radar cross section is assumed to be a good conservative quantity. It is sug-gested that some kinds of in situ formation process should occur in the lower stratosphere.
A simple turbulent energy equation based on semi-empirical turbulence theory was used to parameterize the vertical eddy exchange processes in the atmospheric boundary layer. Numerical experiments with a one-dimensional version of the Techniques Development Laboratory boundary layer model were conducted to evaluate the parameters employed in the turbulent energy equation. These parameters include the ratio αT between eddy diffusivity for heat and that for momentum, and the turbulence length scale l. The study concludes that while specification of l has a great effect on the wind speed, values of αT affect the prediction of temperature substantially. With a proper choice of l and aT, the turbulent energy parameterization scheme may be useful for modeling detailed structures of the boundary layer. Results with four different specification of l and αT are presented and compared with observations. Recommendations are made on the proper specification of l and αT for use in the turbulent energy equation.