A 3-dimensional, anelastic cloud model is applied to the simulation of the July 19, 1981 Cooperative Convective Precipitation Experiment (CCOPE) case study cloud. The model utilizes the bulk water microphysical parameterization technique where number concentrations of cloud ice and snow are taken into account in addition to the mixing ratios of six water species (water vapor, cloud water, cloud ice, rain, snow and graupel/hail). Cloud ice is initiated only by primary nucleation processes (deposition/sorption and heterogeneous and homogeneous freezing of cloud droplets) in the present model. The timing reference was established between the simulation and observations based on a remarkable change in the rise rates of both the observed and simulated cloud tops, and the model results are compared with the observations as a function of time and space. The general features of the cloud (such as cloud top height, cloud size, arrival time of precipitation at the cloud base, radar first echo, etc.) seem to have been well reproduced. Furthermore, the model cloud simulated quite well the location of hydrometeors with respect to the updraft/downdraft structure, the number concentration of precipitating ice particles, updraft velocity and cloud water content along the King Air's penetrating pass. The main features which are not accurately reproduced are the cloud base height, the rise rate of the cloud top and the radar echo near the ground. The cloud base height is too low, which is attributed to the lack of representativeness in the input data taken from the closest radiosonde sounding, while the too rapid rise rate of the cloud top seems to be attributable to the way in which convection is initiated. The rapid decrease in radar reflectivity of the simulated cloud seems to be attributable to inadequate parameterization for rain and graupel.
Through use of the outgoing longwave radiation (OLR) and 700mb height fields for 1979, an investigation was conducted of the variation of the atmospheric circulation over Asia and the western Pacific associated with the 30-60 day variation of the Indian summer monsoon. Results of the analysis of the OLR indicate that convection over the northwestern Pacific along 15°N is active slightly before the active phase of the Indian monsoon over central India, while there is a decrease in convective activity prior to the break phase. This active convection (between 140°E and 150°E) propagates southward with a period of 40 days from 15°N to the equatorial zone. The cloudiness to the north of the Tibetan Plateau (around 55°N, 75°E) and over the subtropical high region (southeast of Japan) reaches a minimum slightly before the active phase of the Indian monsoon. Conversely, cloudiness reaches amaximum prior to the break phase of the Indian monsoon. From the results of the analysis of the 700mb geopotential heights, it was found that to the north of the Tibetan Plateau the maximum cloudiness corresponds to the stagnation of the trough while the minimum cloudiness is associated with the development of the ridge. The position of the subtropical high shifts northward (along 25-30°N) slightly before the active phase of the Indian monsoon and shifts southward (along 10-20°N) slightly prior to the break phase. The phase of the 30-60 day variation to the north of the Tibetan Plateau precedes that of the subtropical high region. These results suggest a close association among the 30-60 day variation of the Indian monsoon, the convection over the western Pacific, the westerly-wave movements over the Asian continent and the subtropical high in the western Pacific. These large scale interactions affect the weather regime around Japan during the Baiu season. The meridional position of the Baiu front (east of 130°E) around Japan shifts from 40°N, when the subtropical high shifts northward, to 30°N when the subtropical high shifts southward. This is due to the variation of the meridional position of the subtropical high, which is associated with the 40 day variation of the monsoon. The meridional position of the Mei-yu front (west of 130°E), in contrast, does not exhibit a cyclical variation.
The internal gravity wave drag formulations developed by Palmer et al. and Lindzen have been included in a diurnally varying model of the middle atmosphere, which extends from the surface to 100 km. The combined impact of these formulations was to improve further the agreement with observation of the model mean zonal wind and temperature distributions. In particular a 20 K warming of the lower stratosphere of the winter hemisphere was obtained, which considerably rectified the previous cold bias. The Palmer et al. scheme is based on orographically excited gravity waves, and it is possible to follow the induced gravity wave stress from the surface to the lower thermosphere in the model. Results for a number of geographical regions show quite diverse vertical profiles, with the wave stress being usually absorbed in the lower stratosphere and mesosphere. However, in some cases absorption only occurred in the mesosphere. Marked diurnal and semi-diurnal temporal fluctuations were obtained in the wave stress at high levels, forced by the tidal variations at levels where the wave stress was partially absorbed. Thus gravity wave-tidal interactions were readily simulated in the model. The major limitation of this scheme was found to be the lack of wave drag generated over the extensive oceanic areas, particularly in the southern hemisphere. The Lindzen scheme only identifies a breaking height for the upwards propagating gravity waves, usually in the upper mesosphere, and thus no information is obtainable on the wave stress profile at lower levels. This scheme does have the advantage of operating at all points of the Earth's surface, but at a given model point it was found to be highly intermittent temporally. The breaking height was extremely sensitive to the zonal wind profile. The diurnal variation of this wind profile, owing to tidal influences, was extremely high in the mesosphere and lower thermosphere of the model. This resulted in a corresponding variation in the gravity wave induced drag, indicating that gravity wave-tidal interactions were also simulated with the Lindzen scheme.
For a two-layered fluid with l2/l1=0.4 (l1=N1/U;l2=N2/U; N1 and N2 are Brunt-Vaisala frequencies of the lower and upper layers, respectively; U, horizontal wind speed), the dependence of non-linear aspects of the flow past a two-dimensional bell-shaped mountain on (l1 h, l1 D) (h: the height of a mountain; D: the depth of the lower layer) is numerically examined, with special emphasis on high-drag states and foehns, using a 2-dimensional non-hydrostatic model. The flow in a high-drag state is generally characterized by the large downward displacement of the air and strong downslope wind with hydraulic jump on the lee side. For π/2<l1D<2π, the transition to a high-drag state occurs when l1h is larger than the critical value (l1h)c. For π/2<l1D<3π/2, the critical value of l1 h for the transition to a high-drag state increases as l1D increases. This is qualitatively similar to the case of l2/l1=0 predicted exactly by Smith's theory (1985). For π/2<l1D<π, transition to a high-drag state is not necessarily accompanied by a wave-induced critical layer. On the other hand, for 3π/2<l1D<2π, transition to a high-drag state is accompanied by a wave-induced critical layer formed in the lower layer. A significant foehn occurs when the flow evolves into a high-drag state with hydraulic jump. For π/2<l1D<3π/2, both the critical value of l1h for the occurrence of a foehn and the coefficient of the linear dependence of a foehn index (a non-dimensionalized potential temperature rise on the lee slope) on the inverse Froude number l1h increase as l1D increases. The critical lh for the occurrence of a foehn predicted by a steady linear theory in θ-coordinates (Smith, 1988) is found to be much larger than that obtained by non-linear numerical solutions.
Operational 15-day forecasts with the global prediction model of the Japan Meteorological Agency (JMA) are utilized to investigate the relationship between the eastward-propagating 30-60 day modes in the tropics and the predictability of the northern hemisphere extratropical circulation for the period from March 1988 to February 1989. It is found that the global model successfully predicts the eastward-propagating 30-60 day modes both in amplitude and phase when they are significant in the observation. There is a tendency that the forecast skill of the 10-day mean 500 mb geopotential height north of 20°N for days 6-15 becomes relatively high in such cases during the warm season (May-October). In these skilful forecasts, wavelike height anomaly patterns are found in middle latitudes in 500 mb maps of the analysis and the forecast, and wave propagations from low latitudes are seen along these wave-like patterns. The above results indicate that when the tropical 30-60 day oscillation is well predicted the skill of extended-range forecast for the northern hemisphere tends to increase due to a successful prediction of the influence of the 30-60 day modes on the extratropical circulation.
Turbulent fluxes have been determined by the eddy correlation method over the open sea area south of Japan as a part of WCRP/OMLET. Velocity components were corrected for the ship motions and turbulent fluxes of momentum, heat and water vapor were evaluated. It was almost neutral to slightly unstable, and the sensible heat fluxes were as small as 60W/m2. While the latent heat fluxes showed a maximum value of about 200W/m2. Based on the direct measurements of turbulent fluxes, the bulk transfer coefficients were determined as CD=1.04×10-3, CH=1.79×10-3 and CE=0.94×10-3 in the low wind speed region. The drag coefficient, CD increase with increasing mean wind speed and the value is almost consistent with the previous studies. A notable difference is observed in the behaviors of CH and CE, particularly in the high wind speed region.
The global scale predictive tracer model designed to simulate atmospheric transport processes as well as a tropospheric and stratospheric chemistry is described. The model consists of a global spectral model and a set of the mass conservation equations coupled through the chemistry terms. The model was tested in both diagnostic and fully predictive mode simulating the hemispheric scale dispersion of radionuclides from the Chernobyl accident. The analysis of the predictive run results indicate that the model is able to predict the large scale mass distribution for periods up to six days when started from a sufficiently accurate initial condition. An extended prediction of the hemispheric tracer distribution was performed up to 24 days; the tracer field in these experiments was updated every 6 days using an initial condition provided by a parallel run of the diagnostic model.
Taking into account the intermittent gust motion within a canopy layer and the influence of both the canopy elements and the underlying ground surface on the turbulent motion within the layer, a new mixing-length model was developed to simulate the exchange of momentum, sensible heat and water vapor between the atmosphere and the vegetated surfaces. It was found that traditional models overestimate the mixing length when the canopy density is neither dense nor sparse, since these models assume that the mixing length is limited by either the canopy elements or the height from the underlying ground surface. According to the present model, for the case of a vertically uniform canopy, the mixing length within the canopy layer is approximately equal to kz (κ : von Karman constant, z : height) near the underlying ground surface. It remains approximately constant far enough from the ground surface, decreasing gradually as the canopy density increases. The validity of the model was determined by comparing model results with observed data.
Intraseasonal variations of upper and lower-level zonal winds, outgoing longwave radiation (OLR) and globally averaged angular momentum (GAM) for northern summers of 1977-1984 are studied using lag correlation statistics. The role of surface frictional drag on atmospheric angular momentum is also investigated by studying the temporal and spatial distribution of surface wind stress in the tropics and its relationship with zonal wind anomalies. It is shown that the 30-60 day GAM fluctuation is accompanied by zonal propagation of convection and 850 mb zonal wind anomalies in the tropical belt. On the other hand, the associated 250 mb zonal wind and surface wind stress anomalies exhibit pronounced standing oscillations. The magnitude of wind stress anomaly is strongly influenced by the climatological zonal wind in the tropics. Large wind stress anomalies are only found in the tropical central Pacific where surface easterly from the climatological Walker circulation is strong and persistent. Therefore, although the tropical lower-level wind anomalies associated with the 30-60 day oscillation propagate continually from the Indian Ocean to the eastern Pacific, large momentum exchange between the earth and atmosphere only occurs at the time when the low level wind anomalies are in phase with climatological easterlies in tropical central Pacific. From the present results, it is suggested that momentum exchange between the lower and upper troposphere may occur in regions of active convection via vertical momentum transport. The tropical central Pacific plays a key role in the linking the atmosphere and earth through angular momentum exchange in intraseasonal time scales.
The tropospheric circulation in the Southern Hemisphere has some remarkable features such as a deep circumpolar trough all the year round, a double-jet in winter months and large semi-annual components in the fields of sea-level pressure and zonal wind. Performance of a 12-year integration with the Japan Meteorological Research Institute general circulation model is presented and compared with the observation for the period 1979-1987. The simulated meridional temperature gradient in July has two distinctive maxima, one at 30°S in the upper troposphere and the other at 60°S in the lower troposphere, and the baroclinity is not zonally uniform. The simulated zonal winds at 500mb show two jets in the Pacific sector, one at 30°S and the other at 60°S, and only one jet in the Atlantic/Indian sector, in accordance with observation. The double-jet structure is somewhat obscured when taking the zonal mean. Quasi-stationary eddy fields with zonal wavenumber 1 at 50°-70°S are associated with this zonal asymmetry. Even in the Pacific sector two strong baroclinic zones occur only from May to October, while there is one during the rest of the year. Between 50°S and 60°S, baroclinity becomes strong during spring and fall, and semi-annual oscillations are found in zonal wind and sea-level pressure. A good simulation of the seasonal cycle of the Antarctic temperature field such as a rapid cooling in autumn of the Antarctic lower troposphere, lack of a well defined temperature minimum (the coreless winter) and coldest atmosphere in early spring, is crucial to a successful simulation of the semi-annual oscillations and the winter double-jet structure.
Non-linear normal mode initialization (NNMI) proposed by Machenhauer (1977) is applied to the JMA spectral limited area models. The normal modes of the models are constructed using a constant map factor and Coriolis parameter, which was first performed by Briere (1982) for a grid model. The initialization procedure is carried out in spectral space. The NNMI is designed to preserve the prescribed boundary conditions of the model. By introducing a cut-off period, gravity waves with very low frequency are excluded from the procedure. The initialization using the adiabatic time tendencies successfully reduced the large time tendencies of the gravity wave modes of the first five vertical modes. The changes of the initial states by the NNMI with a short cut-off period of 6 hours were very small, while the NNMI with a longer cut-off period produced larger initial changes in the baroclinic zone. The NNMI eliminated high frequency oscillations in the subsequent forecast.
The occurrence of transient eddies having a time scale shorter than 10 days has been presented for the 850mb and 500mb height fields over a period of eleven years (1969 through 1979) for the Northern Hemisphere middle latitudes. The population of both cyclones and anticyclones contained in disturbances with a duration of longer than 3 days decreased about 1.5% per year. The intensity of the disturbances also decreased during these years. The transport of sensible heat by transient eddies decreased while the total eddy transport of sensible heat remained unchanged. The relation between eddy heat flux and radiative forcing is discussed.