The MU radar observations were carried out for continuous five days each in September 1985 and 1986 to reveal characteristics of internal gravity waves in the upper troposphere and lower stratosphere. Hodograph analysis and rotary spectral analyses in the vertical direction and time series of the data lead to the following results; (i) Upward propagating and saturating internal inertial gravity waves (IIGWs) are evidently observed through out all the observations in the lower region of the stratosphere lying above the tropopause. The waves are quasi-monochromatic, and representative wave parameters are a horizontal scale of 200 km (ranging from 110 to 330 km), a vertical wavelength of 1 to 2.5 km, a wave amplitude of several ms -1, and a horizontal phase velocity of 3 ms-1. The vertical wavelength and wave amplitude decrease with increasing height and the waves disappear in the vicinity of the Jones' critical level around 19 km. (ii) Another kind of upward- propagating inertial gravity wave, an inertial wave whose period is very close to inertial period is observed at the zonal mean wind turning level, separately from IIGWs just above the tropopause. The horizontal and vertical wavelengths of the wave are about 1000 km and 2 km, respectively, and the wave amplitude about 3 ms-1. We discuss propagation and origin of IIGWs and the inertial wave based on the wave parameters obtained from analyses. From an estimation of the group velocity of IIGWs, it is indicated that IIGWs are advected by the zonal mean wind and then the observed waves around a height of 17 km at the MU radar site (34°51' N, 136°06' E) are likely to have originated more than 1000 km west of the site. Ray tracing of the wave packet on a meridional plane shows that the inertial wave may have been propagated from the tropical upper troposphere over a period of about a week.
Through a quantitative analysis of tropical rainbelts, determination of the critical behavior of the rainbelt as to its position and/or intensity is found. The behavior is related to the rainfall variability in the semi-arid regions of tropical Africa, the Sahel and northern Kalahari regions. Two parameters describing the rainbelt are used; its center of gravity and total rainfall. These parameters indicate the center and intensity of the connective activity of the ITCZ, respectively. The reduction in total rainfall of the rainbelt during the northern summer solstice season is the critical factor determining the continuous reduction of the Sahelian rainfall during the period from the early 1960's to the mid-1980's. On the other hand, the northward advance and increase in the total rainfall of the rainbelt are equally important for the wet conditions of the Sahel during the 1950's. For northern Kalahari, the reduction in total rainfall of the rainbelt during the southern summer solstice season is also responsible for the below-normal rainfall during the period between the late 1950's and the early 1970's. The southward advance and increase in the total rainfall of the rainbelt resulted in the wet conditions during the 1950's. These results demonstrate that the poleward expansion of the tropical annual rainfall zone during the wet period of the 1950's (Shinoda, 1989) was caused by the concurrent poleward advance and increase in total rainfall of the rainbelt during the northern and southern summer solstice seasons. The equatorward contraction of the rainfall zone after this period (Shinoda, 1989) resulted from the reduction in the total rainfall of the rainbelt for each season.
A technique originally devised by Adler and Negri (1988) for estimating the convective and stratiform precipitation areas and amounts of mesoscale convective cloud systems from infrared satellite imagery has been applied to a large cloud cluster observed over the South China Sea during the Winter Monsoon Experiment (WMONEX). The technique was modified to obtain agreement with a previous analysis of ground-based radar data obtained in a limited part of the cluster. The modifications included altering the method for identifying convective cells in the satellite data, accounting for the extremely cold cloud tops characteristic of the WMONEX region and modifying the threshold infrared temperature for the boundary of the stratiform rain area. After obtaining agreement with the radar analysis over a limited portion of the cluster's area and lifetime, the satellite technique was applied to the entire cluster over its full lifetime. The locations of the convective and stratiform rain areas found in this extended analysis were consistent with WMONEX research ship and aircraft data of the cluster. The evolution of total convective and stratiform precipitation indicated by the satellite analysis over the cluster's lifetime was qualitatively consistent with previous radar analyses of other equatorial cloud clusters, except that this cluster appears to have developed a particularly strong stratiform component during its mature stage. The successful application of the Adler and Negri technique to the WMONEX cloud cluster provides encouragement for the use of this method; to develop satellite-based climatologies of the convective-stratiform internal structure of cloud clusters over large regions of the tropics where radar data are not typically available. It also appears that this technique will be a useful complement to future spaceborne precipitation-measuring radar systems.
The characteristics of the upper (200 mb) and the lower (850 mb) tropospheric wind fields for the period between July and October 1980 over the tropical western Pacific are analyzed, focusing on typhoon formation. GMS satellite wind data and rawinsonde data are used to produce the grid point wind fields. The MASCON model (Dickerson, 1978) was used to adjust the interpolated lower wind fields. Most of the typhoons and tropical storms that occurred during 1980 originated from easterly wave disturbances in the ITCZ. The easterly wave disturbances move in a westward direction before recurving and propagating north and northeastward along the region of large high cloud amounts in which they develop into storm intensity. The origin of the wave disturbances is traced back to as far as near or east of the international date line. The intensity of a given wave disturbance in its early stage is not related to the intensity at its most developed stage. Whether the disturbance develops to tropical storm intensity or not is related to the large scale wind fields surrounding the disturbance. Development of easterly wave disturbances into tropical depressions and storms occurs around the area where the easterlies from the east are confluent with both the westerlies from the west and the southerly winds form the southern hemisphere. This confluence area moves eastward and westward over a period of 10 to 30 days and most of the typhoons and tropical storms appear when the westerly wind region expands eastward.
Observations of atmospheric ozone and nitrous oxide were performed with a laser heterodyne spectrometer using a tunable diode laser as a local oscillator at Sendai (38°15'N, 140°51'E), Japan from November 1988 to June 1989. The absorption line spectra measured with ultra high resolution of 0.0013 or 0.0027 cm-1 (40 or 80 MHz) were inverted by a conventional inversion method to retrieve vertical distribution profiles. The observed altitude range is from 5 to 30 km for ozone and from 1 to 35 km for nitrous oxide and the vertical resolution is∼5km. Total column densities were also obtained by integrating the vertical profiles. Errors in the retrieved mixing ratio profiles were estimated by computer simulations for all the plausible causes: ≤10% arising from uncertainties in line parameters, ∼3% from random noise in the spectra when SNR=500, and ≤3% from difference between model and actual temperature profiles. For total column densities, the random error is ≤3%, while systematic error due to uncertainty in the line strength is ≤10%. The retrieved vertical profiles of ozone from January to June 1989 show seasonal and day-to-day variations usually seen in middle latitudes. The total column densities were compared with those measured by Dobson spectrophotometers at Tateno and Sapporo. The result is consistent with the latitudinal dependence of total column density of ozone. The vertical profiles of nitrous oxide were also obtained by the same procedure as for ozone. The result for nitrous oxide agrees with the data obtained by a balloon observation at Sanriku (39°09'N, 141°49'E).
Long-term variation of the mean zonal wind in the upper stratosphere in December is investigated using NMC geopotential height data from 1979 to 1987. The period of analysis is anteriorly extended until 1973 by combining the results of various investigators. It was found that the mean zonal wind at 1 mb, 40°N shows outstanding long-term variation of the order of 10 years. Similar variation was also found in the rocket measurement of east-west wind component at Ryori (39°N, 142°E). It is shown that the long-term variation of the mean zonal wind has a good correlation with the solar activity, ozone partial pressure and solar UV index.