The responses of an expanded hierarchy of barotropic vorticity models to the anomalous tropical divergence associated to the El Nino event of February 1983 are analyzed. The discussion is focused on questions related to the interpretation of the results based on similar models by other authors. The main points are : (1) the importance of the vorticity advection by the divergent component of the wind for the extratropical response in an ENSO event, (2) the reason of the apparent success of more simple models which do not include this term, (3) the role of the upper level convergence near Indonesia for the development of the PNA pattern and (4) the relative insensitivity of AGCM's extratropical response to the longitudinal position of the equatorial convection along the Pacific Ocean, reported by Geisler et al. This paper emphasizes the major drawback of the barotropic model : the need for a correct specification of the divergence/ convergence associated with an anomalous tropical heat source, in view of the fact that subtropical anomalies in the divergence field, induced by tropical heating, make significant contributions to the midlatitude response.
Large scale relationships between surface heat flux (defined as the sum of sensible and latent heat fluxes at sea surface) and the atmospheric circulation in the Northern Hemisphere during the northern winter have been examined, using the singular value decomposition (SVD) analysis. The dominant heat flux anomaly pattern in the North Pacific has large amplitude in the western and the central part of the subtropical gyre, and the one in the North Atlantic exhibits a north-south seesaw in the western Atlantic. These patterns are closely related to the dominant atmospheric circulation anomaly patterns known as the teleconnection patterns, i. e., the dominant surface heat flux pattern in the North Pacific is related to the Pacific/North American (PNA) teleconnection pattern and the one in the North Atlantic is tied to the Western Atlantic (WA) pattern (Wallace and Gutzler, 1981). Variations with a time scale of a few years are dominant in time coefficients of the leading SVD modes for both oceans. Decadal scale or trend-like variation is also seen in those for the North Pacific but is not apparent for the North Atlantic. We also investigated the relation between the surface heat flux and the time rate of change of sea surface temperature (SST tendency). Local correlation coefficients between them are generally high but not close to unity. This suggests that the SST tendency is determined not only by the surface heat flux but also by other processes such as horizontal temperature advection and vertical mixing in the upper ocean. The spatial patterns of the first SVD mode for the surface heat flux and the SST tendency in each ocean are quite similar. They also resemble the heat flux patterns of the first SVD mode paired with atmospheric variables such as 500 hPa height in each ocean. This indicates that the dominant pattern of the SST tendency in each ocean is organized by the dominant large scale atmospheric circulation pattern such as the PNA teleconnection pattern.
Seasonal variations of large-scale convective activity and wind over the western Pacific are examined using Geostationary Meteorological Satellite infrared equivalent blackbody temperature (TBB) and Euro-pean Center for Medium range Weather Forecast (ECMWF) global analyses over a 10-year period from 1980 to 1989. In particular, this study describes an abrupt northward shift of large-scale convective activity over the western Pacific around 20°N, 150°E in late July. The enhanced convective activity is coincident with strong cyclonic circulation there which induces westerlies to the south of the cyclone and easterlies to the north of it. It is emphasized that this strong cyclonic circulation appears suddenly over the subtropical western Pacific region. Monsoon westerlies to the west of 110°E are not similary accelerated at the same time, indicating that this abrupt change is independent of the Asian monsoon system. To the north, an anticyclonic circulation is generated, which corresponds to the withdrawal of the Baiu season over Japan. Furthermore, this abrupt northward shift of large-scale convective activity is shown to be associated with tropical cyclone activity. In the mid latitudes, geopotential height pattern between pre- and post-northward shifts of the large-scale convective activity in late July exhibit equivalent barotropic vertical structure, suggesting the Rossby-wave propagation emanating northeastward from the enhanced convective region around 20°N, 140°E (western Pacific) to as far north as 60°N, 180° (Bering Sea). Another feature is that the seasonal increase of sea surface temperature (SST) over the key area (20°N, 150°E) precedes abrupt convective enhancement by about 20 days, exceeding 29°C in early July. It is inferred that the northeastward extension of the warm SST tongue is intimately associated with the enhanced convection in late July. This result suggests that SST warming is not a sufficient condition but certainly one important ingredient for the abrupt northward shift of convections.
Aircraft observations of a cloud-topped boundary layer were performed during cold air outbreaks in the "water extended cloud experiment" of the "Western North-Pacific Cloud-Radiation Experiment". A three-dimensional sonic anemometer was used to measure the three components of wind velocity relative to the aircraft. The turbulent heat flux due to convection in the sub-cloud layer decreased with height, and the buoyancy flux became negative in the upper part of the sub-cloud layer. It is suggested that the motions in the upper part of the sub-cloud layer are maintained by convection in the cloud layer.
A transient response experiment to the gradual increase in atmospheric CO2 concentration at a compound rate of 1 %/yr has been performed with a coupled atmosphere-ocean general circulation model (CGCM) developed at the Meteorological Research Institute (MRI). The model is characterized by two aspects ; one is a relatively high resolution of the oceanic part in the low latitudes to simulate El Nino phenomena, and the other is an elaborate sea ice model to simulate seasonal variation of sea ice coverage and thickness. Time integration has been performed up to 70 years over which the CO2 concentration doubles. The globally averaged surface air temperature increases 1. 6°C during this period. Atmospheric response to the CO2 increase is slow in the Southern Hemisphere and over oceanic areas. However, the surface air temperature increase in the high latitudes in the Northern Hemisphere is not dominant up to the year 50. This speed of CO2-induced warming is affected by interdecadal variation of sea ice found both in the transient and in the control runs. It is also suggested that leads in sea ice act as a strong negative feedback on changes in sea ice volume, affecting the timing of the warming. Analysis of sea surface temperature shows that the dominant air-sea coupled mode in the model is very close to what is observed. This mode shows interannual variations in the Pacific with a dominant period of about 6 years, which is close to the typical time scale of El Nino. It also shows variations of interdecadal time scales, with implication of predictability for a few decades.
Density currents with cold air circulation are analytically investigated in a 2-dimensional model. Both the propagation speed and depth of density currents are shown to be unaffected by the direction of the cold air circulation for any vorticity distribution. In particular, the propagation speed and depth are explicitly calculated for the case of small vorticities of the cold and warm air masses. The result shows that they increase with the cold air vorticity irrespective of its sign to the leading order approximation. Consistent with the previously published results, they increase also with the warm air vorticity when the vertical integral of the far upstream warm air velocity relative to the ground value is positive.
The effect of non-uniformity of vertical potential temperature gradient on downslope windstorms in a uniform environmental flow is analytically investigated by employing a steady 2-dimensional hydrostatic model. Both the potential temperature and a vorticity-like quantity become functions of the stream function. Their functional forms are chosen so that the environmental flow has a uniform horizonal velocity and non-uniform potential temperature gradient. Only the case of small parabolic deviation from uniform potential temperature gradient with the average value fixed is examined. The calculation shows the following. For a prescribed atmospheric height, the concentration of the potential temperature gradient in the upper (lower) layer increases (decreases) both the downslope wind speed and required mountain height. This result, compared with the case of a uniform increase of potential temperature gradient, implies that the upper layer stability is more effective for downslope windstorms.