We conduct numerical experiments of the quasi-biennial oscillation (QBO) using a 2-dimensionalmodel. An oscillation with period of about 600 days is obtained. The simulated oscillation is verysimilar to the observed QBO in many respects: (1) The faster downward propagation of the westerly wind than that of the easterly is obtained. (2) The westerly wind appears at first on the equator but the easterly spreads from the mid-latitude region. (3) The meridional extent of the westerly wind at the initial onset in the upper layer is very narrow. We analyze the effects of wave transience, change of zonal phase velocity by wave self-acceleration mechanism and wave breaking in the model. The effects except wave self-acceleration are crucial in the present model. The dependence of the results on the magnitude of the dissipation coefficients is examined. We have conducted some numerical experiments with different initial conditions. However, the same steady state solution is always obtained when the magnitude of the dissipation is large. In this case, it seems that no multiple equilibrium exists in spite of the nonlinearity of the system.
The effect of the horizontal structure of midlatitude oceanic heating on the stationary atmospheric response is examined by means of two simple two level quasigeostrophic β-plane channel models. Solutions are obtained for three non-periodic zonal heating structures (line source, segmented cosine, and segmented sine), and these are compared with the results for a continuously stratified model. Little difference is observed between the solutions for these two different models. There are two cases which emerge in obtaining analytic solutions. In case 1, for large meridional wavenumbers, there exists a large local response and a constant downstream response. In case 2, for small meridional wavenumbers, the far field response is now sinusoidal. A critical wavenumber separating these two cases is obtained. The effect of oceanic heating on the atmosphere over the Kuroshio region is examined in an attempt to explain the large correlations observed between winter Kuroshio oceanic heat flux anomalies, and the winter atmospheric surface pressure and 500 & 700mb geopotential heights, both upstream and downstream of the heating region. It is found that the model response is consistent with the observed correlations. When western North Pacific heating and eastern North Pacific cooling are introduced into the model, a large low pressure response is observed over the central North Pacific. This feature is in excellent agreement with the observed correlations.
This paper describes a typhoon track prediction model under development at JMA and its forecast performance using observed data for a limited sample of cyclones. The experimental model uses primitive equations in σ coordinate. The model utilizes a limited area grid and is nested in a one-way sense to the forecasts of a hemispheric model. It has a uniform grid system with a horizontal resolution of about 50km and 8 vertical levels and covers the area of 4000km×4000km. The cumulus convective process is parameterized using Kuo's framework. The numerical schemes are specially tuned to realistically maintain the model typhoon. The initial condition consists of a well-formed model typhoon superposed upon a large-scale objective analysis. This model predicts the central pressures and movements of several typhoons observed in 1985 with a fairly good skill. In addition, this model simulates a complex rainfall distribution under the existence of a strong interaction between a typhoon and Baiu front quite reasonably as compared with satellite cloud pictures. Some impact studies suggest that particular attention should be paid for the accuracy of the manual analyses and the performance of the hemispheric model (background model), because they have large effects on the track performance of this model.
An observational study is made of low-frequency variations of the general circulation in the Southern Hemisphere troposphere with the aid of daily data analyzed at the National Meteorological Center (NMC). In the winter of 1983 there exist two typical patterns of the zonal mean geostrophic wind in the meridional plane. We name them the single jet regime and the double jet regime, each of which persisted with a characteristic duration of a month and appeared twice alternatively. In the single jet regime, the subtropical jet at the tropopause level is strong and the stationary wave of zonal wavenumber (WN=)3 has a larger amplitude than WN=2. On the other hand, in the double jet regime the subtropical jet is weak and the polar night jet extends down to the surface at about 55°S. Dominant stationary planetary waves are WN=1 and 2. These two regimes are also observed in the three years 1980-1982. In terms of the quasi-geostrophic Eliassen-Palm (E-P) flux diagnostics, the vertical component of the E-P flux, which is proportional to the horizontal eddy heat flux, is larger around the mid-latitude troposphere in the single jet regime than in the double jet regime. Consequently, the convergence of the E-P flux, which acts as a wave drag on the mean westerly wind, is larger around the mid- and high latitude tropopause level in the single jet regime than in the double jet regime. The larger wave drag tends to keep the weaker westerlies in the single jet regime. An empirical orthogonal function (EOF) analysis is made to define an index of the zonal mean wind of two dimension The first component of EOF, which represents over 40% of the total variance, corresponds to the variations between the two regimes. The amplitude of this component is used as a kind of 'zonal index'. Time variation of the zonal index is not considered an internal oscillation with an intrinsic frequency because the 'period', which is not an appropriate concept for the present phenomenon, is different in each year. The transition between the regimes is rather rapid compared with the duration of each regime. No clear relation is found between the regime transitions and the annual variation. These observational evidences support an interpretation that this low-frequency variation of the zonal index is the transition between multiple planetary flow regimes in the Southern Hemisphere.
The 10-20 day variations in the northern hemisphere winter-time wind field in the upper troposphere have been examined using FGGE data. Previous study by Pan and Zhou (1985) has demonstrated the importance of the upper tropospheric long-synoptic scale waves in the modulation of the cold air outbreak over the winter monsoon domain. Results of this study indicate that these longsynoptic scale waves are part of a hemispheric wavetrain that propagates slowly eastward. Coherent wave movement can be traced from the Atlantic Ocean eastward to the central Pacific Ocean where the waves move equatorward. Teleconnection type response in the western Pacific Ocean is observed to be strong when a cyclonic or an anticyclonic center of the wave moves near the dateline. Singlelevel regional kinetic energy budget has been examined for these long-synoptic waves and, to a first order-of-magnitude approximation, the maintenance of the long-synoptic waves can be viewed as a balance of the flux-divergence by pressure-work and the barotropic energy conversion to mean flow.
In winter, 1984, snow cloud bands (Sakakibara et al. 1987) were observed over the Japan sea off the coast of Kanazawa. They lasted about 90 minutes, and the orientation of the bands was almost perpendicular to the direction of mean wind. 2-dimensional numerical simulation of the snow band is performed, using a non-hydrostatic anelastic model with the bulk parameterization of cloud microphysics including both water (cloud water, rain) and ice phase (cloud ice, snow, graupel). In the early stage of the simulation, a unicellular storm is reproduced, and later, it turns into a multicellular storm. In this paper, the simulated multicellular storm is investigated in detail. It is com-pared with the storm at the unicellular stage, observations, and other squall lines so fax documented. The characteristic features of the multicellular storm and temporal behavior of individual cells which constitute the multicellular storm are as follows. 1) The fields of the multicellular storm averaged on the storm-scale (-15km, as contrasted with the individual cell scale-5km) are similar to that of the unicellular storm in these points; a) upshear tilting of updraft, front-to-back ('front' is defined as the direction from which the low level system-relative flow enters into the storm) systemrelative u-component of updraft branch, and back-to-front system-relative u-component of downdraft branch, b) a cold dome at the surface and corresponding high pressure area, c) a storm-scale warm area and corresponding low pressure area above cold dome. But the horizontal scale and intensity of the fields are larger than those of the unicellular storm. 2) The fluctuations by individual cells are large in w, cloud water, and graupel fields and at the mid level, while those are relatively small in other fields and at the low level. 3) Transformation of the unicellular storm into the multicellular one is related to the faster movement of the gust front than that of the individual cell. 4) A new cell is generated above the gust front intermittently at the time interval of 25 minutes on average but with large variance, and at the distance of about 5km on average but with large variance from the old cell. 5) After the generation of a new cell, the warm air inflow from the low level into the old cell is cut off, and the old cell begins to dissipate and move backward. 6) In the new cell, cloud water and graupel are found, while, in the old cell, little cloud water is found and snow is dominant. The results of simulation agrees well with observations in the following points: 1) multicellular structure, 2) upshear tilting of updraft, front-to-back system-relative u-component of updraft branch and back-to-front system-relative u-component of downdraft branch, 3) a cold dome at the surface, 4) the width and height of the radar echo of the snow band, 5) long lasting property. The features of the snow cloud band observed and simulated indicate that the snow band is similar to the squall line, although the strength and scale of the cloud band and the environmental thermal stratification are different.
Collision experiments were conducted with model snowflakes suspended on the end of a needle in a vertical wind tunnel to examine the effects of surface roughness and porosity of snowflakes on their collection efficiency for cloud droplets. The snowflake models were disks with small glass beads attached to their surface (roughness) or rectangular holes evenly spaced on their surface (porosity). The models were exposed to an airflow carrying saline water droplets or spherical glass particles. The effects of roughness and porosity were deduced from counting the numbers of captured particles on the model surface or from analyzing the motion of the particles around the models. The results showed that the collection efficiency was enhanced by these effects, especially in a small inertia region. Very small particles, which could not be scavenged on the basis of impaction theory, were captured by the models owing to these effects. Generally the effect of the porosity was larger than that of the roughness. These results suggest that actual falling snowflakes can gather more cloud droplets in the smaller sizes than might be expected from the traditional impaction theory.
With use of a simple climate model, we discuss analytically the cause of appearance of oscillatory climate state on the earth. Introducing a geographical parameter into the climate model, we investigate a relationship between emergence of oscillatory climate and continental drift. Our results can explain mathematically the observed fact, which gives records during last million years, that it grew colder with a gradual increase of amplitude of climatic oscillations, and can indicate that oscillatory climate had appeared on the earth when the geographical condition had allowed it to exist. Our tentative calculations reveal that the oscillatory climate might have emerged on the earth, when a clear cryosphere had been established on the earth owing to drift of such lands as the Antarctica or Greenland from warm low-latitudes to ice-possible high-latitude regions, and when the past earth's glacial productive capacity could have come to exceed about 70% or so of the present capacity.
This is a preliminary report on a new measurement of the fractal dimension of horizontal cloud pattern over the tropical oceans, so-called intertropical convergence zone (ITCZ). The physical processes of formation of the organized cloud pattern over the ITCZ are still little understood. The purpose of the present work is to cast some light on this problem from a geometrical point of view. A percolation analysis of an infrared satellite imagery is performed for a whole cloud structure of the ITCZ, which represents a strong dependence on the equivalent black body temperature TBB. For -10≤TBB≤10°C, the cloud patterns are self-similar with the fractal dimension D_??_1.5, which is larger than the value determined by Lovejoy (1982) using a different approach for individual clouds. However, for TBB≤-20°C, the cloud patterns do not represent self-similarity.