Interannual variations of the tropical atmosphere in a 20.5 year integration with an atmospheric general circulation model (GCM) forced by the observed near-global (40°S-60°N) sea surface temperatures from September 1969 to February 1990 are presented. Simulated interannual variations in the tropical Pacific related to El Nino/Southern Oscillation (ENSO) are comparable to the observations. An eastward-propagating signal which moves from the Indian Ocean to the central Pacific taking about one and a half year is clearly demonstrated by the 850mb tonal wind not only in the observed data but also in the simulated fields. The annual cycle and interannual variations of surface wind stresses over the tropical Pacific are investigated. The empirical orthogonal function (EOF) analysis of the zonal wind stress anomalies shows that the simulated first arid second EOF modes resemble the observed counterparts both in its spatial structure as well as the coefficient time-series. The eastward propagation of the zonal wind stress anomalies represented by these two modes has been reproduced by the GCM. The magnitude of the simulated zonal wind stress anomalies in the central Pacific is comparable to the observed counterpart. However, the model tends to show smaller zonal wind stresses over the western tropical Pacific during the development stage of ENSO. The magnitude of the simulated meridional wind stress over the equatorial Pacific is much smaller than the observation in its mean annual cycle as well as in its interannual variations, although its spatial structure arid time evolution is very similar to that observed.
An atmospheric general circulation model is integrated, forced by the observed near-global (40°S-60°N) sea surface temperatures (SST) from September 1969 to February 1990. Recurrent patterns of the rnorithly mean Northern Hemisphere atmosphere in the simulated 20 winters are identified and compared with the observations. The simulated first mode, in the zonal wind at 200mb, geopotential height at 500mb, and zonal-mean zonal wind, fluctuates with a time-scale of about 10 years. This decade-scale variation seems to result from the specified SST for this period particularly over the tropical western Pacific northeast of New Guinea. A meridional shift of the East Asian subtropical jet and associated circulation changes contribute to this variation. Surface air temperature variations from around Japan to the extreme northern Pacific are mostly explained by this mode. This mode resembles the observed first and fourth modes of thee geopotential height at 500mb. The model's systematic error of emphasizing variability over the North Pacific more than that over the North Atlantic results in extracting this mode associated with the meridional shift of the East Asian jet as the leading mode. Additionally, a trend of the equatorward shift of the Southern Hemisphere subtropical jet is also simulated related to the trend in the SST.
The transient response of planetary waves to tropical heating is investigated using a global atmospheric model linearized with respect to basic zonal flows which include both vertical and meridional shear. This work focuses examination on the role of baroclinic instability in generating external-mode motions under the influence of a specified anomalous tropical heating. We found that the process of baroclinic instability of zonal flows can be considered as a cooperative action with tropical heating in the generation of external-mode motions which cause global barotropic-type responses. However, the effectiveness of each process in energy generation is different. For example, the efficiency of generation with tropical heating is greatly reduced in the easterlies and also with decreased longitudinal heating scales. By contrast, the efficiency of generation with baroclinic instability has the opposite tendencies. Hence, the two processes are likely complementary to each other. Other information obtained from the present transient study, but not from previous stationary response studies, is the time scale of generating external-mode motions by the two processes. The time scale of response involving tropical heating as measured by the leveling-off of internal-mode energy level is approximately 10 days. On the other hand, the time scale of response that it takes to produce the solutions close to the steady states under stationary heating is in the range of 20 to 40 days under reasonable energy dissipation rates.
The structure of stationary planetary waves in the winter stratosphere is computed by. means of a steady-state hemispheric quasi-geostrophic model with a zonal basic state and lower boundary forcing obtained from climatology. The nonlinear wave solution is found to resemble rather closely the linear one, despite the large wave amplitudes which distort considerably the westerly zonal flow. Zonal wavenumber one is the most affected by the wave-wave interactions, experiencing an increase in amplitude and a decrease in westward phase tilt in the northerly regions of the middle stratosphere. Comparison of the solutions to the corresponding climatological wave structure indicates that the inclusion of the nonlinear terms leads to an improvement of the structure of wavenumber one. An examination of the 5-year January climatological basic state reveals a distinct linear relationship between the zonal streamfunction and the tonal potential vorticity in middle and northerly latitudes. Consequently, the wave-wave interactions are to a first approximation a result of the presence of the model dissipation. Weak dissipation in this region implies only weak interactions, which explains the quasi-linear structure of the solutions.
The ascent rates of individual radar echoes of thunderclouds in the Hokuriku district were investigated in relation to lightning activity. Both CAPPI radar and the sferics direction-finding system were used for this investigation. Multicell-type thunderstorms containing several moderate to strong precipitating domains were observed by means of a conventional 5.7 cm weather radar and 100.5 MHz sferics direction-finding system installed at Komatsu Airport and its periphery. The results are briefly summarized as follows: 1) The first lightning discharge appears about five minutes after the 30 dBZ reflectivity echo exceeds the -20°C temperature level. 2) The mean ascending velocity of echoes just before the initial reception of sferics from them is about the same both in summer and winter. By contrast, in cases without lightning activity, the 20-to-45 dBZ echoes have nearly the same ascending velocity as those with lightning activity in winter, but, in summer, the 20 and 25 dBZ echoes have a higher ascending velocity and the 30 and 35 dBZ echoes and equal or lower velocity than those with lightning activity. 3) In cases of very intense lightning activity, the 20-to-35 dBZ echoes ascend much faster than the 40 and 45 dBZ echoes in summer, while in winter, the 20-to-35 dBZ echoes ascend slowly or remain stationary and the 40 and 45 dBZ echoes ascend very fast. 4) In both summer and winter, the peak of lightning activity is observed when several strong echoes of 45 or 50 dBZ are formed at the -10°C temperature level and descend toward the 0°C temperature level. Takahashi (1984) established a rational thundercloud model through numerical calculation and clarified the characteristic convective and electrical activity of cloud cells in correlation with their life cycle. The present observational results can be interpreted as evidence that Takahashi's model corresponds well with actual thunderclouds observed in the Hokuriku district.
Tropical wave disturbances, which later developed into tropical storms over the equatorial Pacific (11OE-9OW, 3OS-3ON), are analyzed using the satellite and rawinsonde wind data in July and September, 1980. Daily wind fields at upper (200mb) and lower (850mb) levels on grid of 1°×1° are composed and analyzed in relation to tropical storm development. There were 10 tropical storms during these two months in the western Pacific, and 5 of them are analyzed, their initial wave disturbances coming from a far-eastern area in the western hemisphere, except one which appears in higher latitudes. The four tropical storms developed from the easterly wave disturbances, which travel with easterly wave of about 5 day in period along 10N line from the eastern Pacific, around 150W. With the westward movement of easterly wave disturbances, the upper cold lows are cut off from the Mid Pacific Trough in the middle latitudes at about 150W and move westward along 20N. The upper cold low is located north to north-east of the easterly wave disturbances in the early stage, but it is seen in the north-west portion of the easterly wave disturbance in the western Pacific. These tropical disturbances in the western Pacific, as well as those in the eastern Pacific around 110W, develop into tropical storms when they encounter the ascending motion of the zonal circulation cells, which are seen around 140E and 110W. The position and the strength of the tonal circulation cell changes with some periodicity in relation to intraseasonal variation. The subsidence motion is suggested around 140W for the western cell, and the initial wave disturbances of typhoons are first recognized to the west of this subsidence.
To better understand the selection of eastward-propagating equatorial wave-CISK modes simulated by Miyahara (1987), and Lau and Peng (1987), a simple linearized wave-CISK model having one or two vertical modes is utilized. Two types of wave-CISK processes (traditional and positive-only) are adopted. Second-order horizontal viscosity/diffusion is included to obtain the preferred modes with horizontal scales comparable to the equatorial radius of deformation. Without this viscosity/diffusion, the horizontal scale of upward motion becomes infinitesimally small. By changing the beta parameter and vertical heating profiles, which makes up the first and second modes, characteristic features of the modes appearing in the model are theoretically and numerically examined. On a non-rotating plane, excited motions are always symmetric in the zonal direction around the initial impulse. Neither preferential propagation nor selective amplification occurs. Three regimes are obtained: (1)a decaying adjustment mode, (2)a stationary growing mode, and (3)a pair of zonally propagating growing modes having the same properties except for the propagation direction. On the equatorial beta plane, three regimes are similarly obtained, corresponding to those on the non-rotating plane. However, the deviated westward propagation appears and the symmetric features in the east-west direction are destroyed. This deviated westward propagation is produced by the dual roles of the equatorial beta effect for the equatorially symmetric convective modes-rotation and a side wall effect at the equator. In the positive-only model on the equatorial beta plane, drastic changes take place between the pair of oppositely propagating modes in regime (3). The westward-propagating modes sequentially create new modes to the east and form a wave packet, resulting in decaying modes. On the contrary, the single eastward-propagating growing (EG) mode is not changed compared with that on the nonrotating plane. As a result, the EG mode is selected as the preferred mode. This EG mode has a westward-tilting baroclinic structure and is similar to those simulated by Miyahara, and Lau and Peng. The selective amplification of the EG mode is due to two oppositely propagating modes, which are isolated by the positive-only wave-CISK, that induce different linear responses on the equatorial beta plane in the east-west direction.
Two convergence cloud bands over the northern part of the Japan Sea are simulated with a 2-dimensional non-hydrostatic cloud model and the results are compared with special observations performed by the Japan Meteorological Agency. The model covers 600km in the horizontal direction and includes Primorskii, USSR, the northern part of the Japan Sea, and Hokkaido, Japan. We successfully simulated convergence cloud bands nearly 100km off the west coast of Hokkaido and 70-80km off the east coast of Primorskii. The results show that the cold air from the Asian Continent stops at the west convergence cloud band and the land breeze from the Hokkaido also stops at the east convergence cloud band. Experiments with different lower boundary conditions and physical processes are made to examine the effects of initial wind, surface temperature, latent heat release and orography upon the convergence cloud bands. The results show that the pressure gradient induced by the gradient of the sea surface temperature causes the westerly wind between the two convergence cloud bands. Latent heat release is necessary for convection to develop fully as observed. We also find that both the sea-land temperature difference at the coast of Hokkaido and the initial wind in a lower layer are important factors in determining the position of the east convergence cloud band.
The response of the surface mean wind vector to the synoptic pressure gradient was statistically calculated over the region of central Japan. The response was examined as related to the thermal effect, caused by daytime heating and nocturnal cooling in the surface layer. In the process of vector accumulation for a large number of stations, the wind components which originate from local circulations cancel each other. Consequently, in the analysis, it is made possible to extract the surface wind component induced by the synoptic pressure gradient from the actual wind data (this component is termed the SWC). The thermal effect was parameterized by the diurnal component of temperature change (DCTC), using a high-pass filter applied to the AMeDAS temperature data. As a result, a close relationship among the SWC, the synoptic pressure gradient, and the the thermal effect was found. Although the strength of the SWC is directly proportional to the strength of the pressure gradient, the response of the SWC to the pressure gradient can be modified by the DCTC. That is, the response is increased by daytime heating and decreased by nocturnal cooling. Moreover, the response undergoes large change at a critical pressure gradient (1.2mb/100km). The critical gradient is more obvious under cool conditions. It was concluded that the critical gradient separated surface airflow patterns into two categories; airflow patterns dominated by local circulations and airflow patterns mainly controlled by the SWC.
Laser radar observations of cirrus clouds were carried out during the spring of 1987 and the winter of 1987/88 at Nagoya, Japan. In most cases the integrated backscattering coefficient of the clouds ranged from 0.001(/strad) to 0.1(/strad). The maximum value of the backscattering coefficient of the clouds smoothly decreased with temperature below -40°C and -20°C in the spring and winter respectively, and the value of the winter coefficient was smaller than that of the spring in the temperature range below -40°C. The temperature dependence of cloud depth was different in the two seasons, and the minimum depth was observed near -40°C. The lowest cloud temperature was about -75°C in the spring, whereas it was limited to near -50°C in the winter. Double-layered cirrus clouds were frequently seen during the observation period. The upper layers were located near the tropopause, and the height differences of the two layers were about 2-3km. Most of them were observed in the spring, and they sometimes continued for longer than a few hours. The relation between cirrus clouds and jet streams was studied. The results showed that the lower limitation of the winter cloud temperature was due to the facts that a jet stream was usually located over Japan and only the clouds below the jet stream could be seen in the winter; the clouds were frequently far below the tropopause with high wind speed area between; and the tropopause on several occasions was located below the jet stream. The winter cirrus clouds were observed around two types of jet stream. One was in the anticyclonically curved area of a meandering jet stream, and the other was in the vicinity of a strong and straight jet stream. On the basis of these results, we discuss the optical properties and generation mechanisms of the cirrus clouds in the vicinity of a jet stream.
A coherent Doppler lidar has been developed using a hybrid CO2 laser with a low pressure gain section below the threshold. The CO2 laser oscillates in a single longitudinal mode with a modelimiting low pressure gain cell within its cavity. The offset frequency between the hybrid CO2 pulse laser and the local oscillator cw CO2 laser is controlled by monitoring the heterodyne IF beat frequency of the hybrid CO2 laser pulse scattered from lidar mirror surfaces. Configuration for the lidar system is simpler than that for the recently developed coherent lidar with an injection-locked TEA CO2 laser. Lidar performance was evaluated by observing wind vector profiles in the boundary layer. Wind variation caused by a sea breeze front was observed in the evaluation. Despite the lidar performance, much effort is needed to decrease the chirp of the hybrid laser, since the laser has a high frequency chirp within a pulse. Otherwise, it might be preferable to use another laser with much less chirp to improve lidar reliability.
Further numerical experiments are performed to supplement the numerical study of Nagata et al. (1986) on the formation of the convergent cloud band (CCB) over the Japan Sea in winter. The purposes are to evaluate the relative contributions of the three lower boundary forcings (the landsea thermal contrast, the blocking effect of the mesoscale mountains, and the characteristic SST distribution) to the formation of the CCB, and to investigate how the mountains north of the Korea Peninsula produce the low-level convergence zone. Simulated mesoscale fields show that the contributions of each lower boundary forcing to the formation of the CCB are comparable. Convective heating is less important than sensible heating over the northwestern part of the Japan Sea in contrast with over the southern part. We make cross-section analyses and examine horizontal inertial stability of the CCB over the northwestern part of the Japan Sea. The dynamical effect of the mountains is shown to produce a local baroclinic environment, including a zone of near neutrality both to inertial and gravitational stabilities. Surface sensible heating appears to maintain the CCB in this zone. These findings suggest that the CCB is a mesoscale longitudinal roll parallel to the vertical shear vector associated with the local baroclinity.