During the Baiu Front Heavy Rainfall Experiment in 1988, a mesoscale rainband was generated along the Baiu front and heavy rainfall up to 178mm occurred. Based mainly on dual Doppler radar observations, we have analyzed the three-dimensional configuration of radar echo and the circulation in the rainband and will describe its structure, focusing on its kinematical features and on the maintenance process. The mesoscale rainband was generated on 17 July 1988 and was maintained for 7 hours in an environment in which the low-level horizontal temperature gradient was not intense and the thermodynamic instability was between that of the tropics and the midlatitudes. The rainband was 170km long and consisted of a convective region and a trailing stratiform region. The orientation of the rainband was from the northwest to the southeast, being nearly parallel to the vertical wind shear between middle levels and lower levels. While existing convective-cells in the convective region moved along the rainband, new convective cells were successively formed at its southwestern edge where the low-level environmental wind entered the rainband. There were found several specific flows in the rainband: 1) a convective-scale updraft located at the front part of the rainband, 2) a convective-scale downdraft at the heaviest rainfall region, 3) a middle-level rear inflow with dry air from a reflectivity notch, 4) a mesoscale downdraft extending from the rear inflow, and 5) forward and backward divergent flows at the lowest level below the convective-scale downdraft. The lowest divergence flows produced a cold pool that was ∼4°C colder than the environment, and a gust front was formed between the cold pool and the moist, warm southwesterly flow ahead of the rainband. The synoptic-scale dry layer in middle levels behind the rainband had an important role in driving the rainband, as well as the moist warm air at the lowest level ahead of the rainband. Upper sounding data showed that a negative temperature deviation, which was made by evaporative cooling of precipitation particles, existed in the convective-scale downdraft and the mesoscale downdraft. The dry layer associated with the rear inflow reached the lowest level of the convective region. These thermodynamic features are consistent with the kinematical structure deduced from the Doppler radar analysis. The rainband was generated in a midlatidude frontal zone, but in the lower troposphere the difference of thermodynamic conditions between ahead of and behind the rainband was very small, as in the tropics. The rainband could be called a ‘self-maintaining convective system’ as well as tropical or midlatitude squall lines, rather than mesoscale systems as observed along the polar fronts of Western Europe or the Pacific coast of North America.
Some numerical experiments are conducted in order to investigate the performance of the SBC (spectral boundary coupling) method proposed by Kida et al. (1991). The SBC is a new method for long-term integrations of a limited area model for regional climate simulation studies. The modified version of the FLM (the fine-mesh limited area model) which was operationally employed for short-period weather forecasting in the JMA (Japan Meteorological Agency) from 1982 to 1988 is used in present study. The model is nested in the GANL (the JMA operational global analysis data sets) using the one-way nesting method. One-month integrations are performed using (a) the SBC method and (b) a conventional lateral boundary condition (LBC), and these results are compared. In addition to those methods, the boundary relaxation method is used in order to suppress noise development near the lateral boundary. The simulated phase of synoptic-scale disturbances using the conventional LBC method tends to deviate from that of the GANL. However, when the SBC method with a spectral boundary condition of Nc=3 is employed, the phase deviation from the GANL is small during the one-month integration. Two experiments with the spectral boundary conditions of Nc=3 and Nc=1 in the SBC method are compared in order to examine the dependence of the model performance on the specification of the spectral boundary wave number. The phase in the case of Nc=1 deviates from that of the GANL as compared with the Nc=3 case. Statistical verification analyses of the precipitation which are simulated using (a) the SBC method with Nc=3 and (b) the conventional LBC method are conducted. The results show that the SBC method provides improved performance of precipitation in one-month integrations of the regional model.
The characteristics of snowfall and airflow over a low mountain barrier (the Dewa Hills; the height of the ridge is about 0.6km) were investigated using a Döppler radar and rawinsondes. A sequence of snow clouds appeared accompanying a cold-front passage on 3 February 1990. Radar echoes were observed to pass over the Dewa Hills for about five hours, and they had diameters of tens of kilometers. The mixed-layer top was lowered and the wind was accelerated over the low mountain barrier, and the flow was considered to be ‘subcritical’ flow. The lowered mixed-layer top boundary suppressed the cloud- and echo-top height. The echoes contained the cells whose sizes were 5km to 10km. The cells were accelerated by 3m/s when they were over the mountain barrier. The precipitation enhancement occurred over and as far as 10km upwind of the mountain barrier. It is considered that the dominant process of the precipitation enhancement was a seeder-feeder process between the snow clouds and the clouds formed by the orographic lifting. On the lee side, colder air in the lowest layer suppressed descending flow and prevented the precipitation reduction by sublimation. It is considered that aggregation and melting mainly increased reflectivity on the lee side. In this case, the behavior of the snow clouds was strongly affected by the flow over the low mountain barrier, while the mean flow was not significantly modified from ‘subcritical’ flow by the existence of the snow clouds.
Non-linear dynamics of a barotropically unstable circumpolar vortex in a forced-dissipative system is investigated numerically with a high-resolution barotropic model on a spherical domain. Horizontal transport and mixing processes of a passive tracer are also investigated for several types of evolving flow fields with a high-resolution transport model. Vacillating and irregular solutions are obtained within some parameter ranges as well as steady-wave solutions. In the vacillation, planetary waves propagate eastward, changing the horizontal structure periodically. Stepwise transitions from steady-wave solutions to irregular ones via vacillations are found for a sech-type jet which is unstable mainly on the poleward flank of the jet, while no irregular solutions are obtained for a tank-type jet which is unstable on the equatorward flank of the jet within realistic parameter ranges of the stratosphere. Time evolutions of an ideal passive tracer field show that the edge of the circumpolar vortex is so robust that mixing through the edge hardly occurs whether the solution is steady or not. Exceptionally intermittent intrusion and ejection of fluid through the vortex edge occur in the irregular solutions for the sech-type jet.
The quasi-seven-month oscillation (QSO) found in the atmospheric relative angular momentums and the length of day is analyzed in detail by using the Japan Meteorological Agency data. The major source of the QSO is the equatorial tropospheric zonal winds, in particular between the equator and 15°S. However, it seems to have been disturbed or masked by some effects of the ENSOs, behaving as an ENSO-connected intermittent oscillation. Its large amplitude, amounting to one third of that in the stratospheric QBO, results from its phase independency of height. Probably disparate from the QSO, exists the eight-to-nine-month oscillation in the equatorial troposphere stationarily through all the data periods, including the El Niño periods.
An abrupt seasonal change in cloud features around the East Asian cloud band (EA-band) was found in the middle of May or late May in all the years of 1982 to 1987 by using two indexes derived from GMS infrared data, middle- and high-level cloud amount (MHc-amount) and the appearance frequency of the high-level cloud region (Hc-frequency) for a 10- or 15-day period. The change occurred over a short period, less than 5 days (transition period: TP), nearly simultaneously with the onset of the “Southeast Asian summer monsoon” about one month before the onset of the Indian monsoon. The main results are as follows: 1) The MHc-amount and Hc-frequency increase considerably in the coastal region of South China after the TP. The greater part of the Hc-frequency are associated with the deep convective clouds. This change is associated with the enhancement of the low-level southwesterly wind, an decrease in thickness of baroclinic layer below the 500mb level and an increase in specific humidity there and in the region to the south of the EA-band. 2) MHc-amount and Hc-frequency in the EA-band also increase to the east of 130°E after the TP, accompanied by the enhancement of the southwesterly wind in its southern part and the southward shift of a subtropical high pressure area west of 150°E. 3) As for the year-to-year variation, the distribution of the area with large MHc-amount in the EA-band before the TP seems to be greatly influenced by the amount of confluence between the southwesterly wind in the north of the subtropical high pressure area and westerly wind north of the EA-band. 4) MHc-amount over Central China after the TP is about 60% in the year of stronger baroclinicity there, but is about 20% in the year of weaker baroclinicity. 5) Only in 1987, the EA-band and a subtropical high pressure area east of 130°E shifted northward after the TP, which are the opposite characteristics to all the other years.
A new rain-drop scheme named box-Lagrangian rain-drop scheme is developed. In the Eulerian raindrop scheme, since the change of the mixing ratio of rainwater q during unit time is calculated from ∂q/∂t=V∂q/∂z, the time step interval Δt is restricted by the Courant-Friedrichs-Lewy (CFL) condition for rain falling: VΔt/Δz<1, where V is the terminal velocity of rain falling, Δz the vertical grid spacing. In the box-Lagrangian rain-drop scheme, the above numerical constraint can be relaxed by the following method. The bulk of rainwater in a vertical grid box is dropped while keeping V constant during a time step interval, and it is partitioned into grid boxes existing in the space where it is dropped. When Δt is less than the critical value, Δtc, which is calculated from the CFL condition (i. e., Δtc=Δz/V), the box-Lagrangian scheme coincides with the Eulerian scheme. Furthermore, even when Δt is several times larger than Δtc, the box-Lagrangian scheme drops rainwater stably with sufficient accuracy. The box-Lagrangian scheme is effective especially when many vertical layers are employed in the lower part of the model domain to express the atmospheric boundary layer in detail.
Two dominant modes in the northern summer sea surface temperature (SST) variability over a recent 24-year period (1970-1993) are shown together with a corresponding geopotential height pattern at 500hPa. These geopotential height patterns are reproduced by a general circulation model (GCM) when the observed global SST are specified. The first mode has a decadal time scale and is characterized by increasing equatorial central Pacific SST and the Indian Ocean SST with decreasing North Pacific SST. Positive 500hPa geopotential height anomalies in the tropics and negative height anomalies in the North Pacific are associated with this mode. Experiments with different SST configurations show that the tropical Pacific SST is most responsible for these recent decadal-scale atmospheric variations in the northern summer.
The error and the vertical resolution of ozone density retrieval have been estimated as a function of the spectral resolution of the instrument. For a spectral resolution of 0.05cm-1 wavenumber, which is relevant to the IMG mounted on the ADEOS satellite, the tropospheric column density will be retrieved with an error of 10-15% when unknown parameters are restricted to the ozone density and the temperature. It is also suggested that the spectral resolution needs to be at least 0.05cm-1 wavenumber for the purpose of observing the tropospheric ozone without suffering the contamination by the stratospheric ozone.
Tropical West African rainfall data at a one-degree latitude interval within the range 4°N-19°N, averaged over the longitude band of 5°W-5°E, are updated through 1993 by using monthly station rainfall data for the July-September period. Making use of the latitudinal rainfall data, the West African rainbelt indices for 1919-1990 are calculated, (i. e., the latitude of its center of gravity and the total rainfall that occurs over the entire rainbelt zone). The Sahelian droughts for 1980-1984 are primarily associated with weakened convection over the entire West African rainbelt zone, while the southward retreat of the rainbelt plays a secondary role in the drought conditions. The slight increase in Sahelian rainfall for 1988-1989 is related to the near-normal position and total rainfall of the rainbelt. During the late 1980s, the increase in rainfall was widespread over the tropical latitudes of West Africa, while it was more evident south of the Sahel than in the Sahel.