A comparative diagnostic study of upper-air persistent atmospheric events, as simulated by a general circulation model (GCM) and as observed, is presented. We start with an overview of the several theories that attempt to explain such phenomena. Particular emphasis is put on the model approach of Shutts (1983). We next show that the spatial distributions of persistent events is qualitatively similar in the GCM and observational data. The North-Atlantic events are extracted and a rotated empirical orthogonal function (REOF) analysis is done on the resulting data sets. The two REOF sets that are thus obtained are shown to greatly resemble one another. Both explain roughly 50% of their original data's variance. The relationships between the modes within a set are presented, so as to understand their probable combined evolution. The fourth chapter contains an evaluation of Shutt's theory. There, the third chapter's results are used to isolate a particular class of events, namely the strong +ATL2 dipoles. The time-tendencies associated to short time-scale synoptic waves are evaluated, using an E-vectors approach, taking care to distinguish between the onset, mature and demise phases of the events. It seems that these synoptic waves have a significant impact of the average life-cycle of this +ATL2 type of events, whether they be simulated by a GCM or obtained from a NMC set of analyses.

We identify physical processes and localized atmospheric structures that modulate the global-scale atmospheric available potential energy (APE). A 1979--95 time series of Northern Hemisphere APE is derived from the National Centers for Environmental Prediction (NCEP) reanalysis. We discern an average three-day cycle in the APE depletion rate dA/dt with a time series of intraseasonal (1.6--180-day) fluctuations. Synoptic-scale APE depletion events are falls and subsequent rises in dA/dt, and are classified as APE collapses when exceeding climatologically-derived thresholds of APE fall and dA/dt. Three associated regional baroclinic waves are observed. West Pacific warm surges (Type A) accompany cyclogenesis over Japan, anticyclogenesis over the west-central North Pacific, and shallow East Asian cold surges. Bering warm surges (Type B) comprise intense southerly flow across the Bering Strait, cyclogenesis near the Kamchatka Peninsula, and intense anticyclogenesis over Alaska. Atlantic Canada warm surges (Type C) include onshore warm flow and eastern North American storm tracks. / January/February 1989 brought the largest vacillation- and synoptic-scale APE collapses of the study period. A pronounced Madden-Julian Oscillation (MJO) modulated equatorial convection over the Indian Ocean during the APE build-up. The initial APE collapse coincided with Type A development and intensifying MJO convection shifting to the West Pacific warm pool, which was prominent because of the 1988--89 La Nina. Extensive large-scale ascent amplified the thermally direct Hadley cell circulation, with descent strong over East Asia. The second collapse was associated with Type B development characterized by a record-breaking Alaska anticyclone. / Factors explaining lacking precision in energy budget calculations include missing o values above 100 hPa, limited representation of subgrid processes, and limited temporal resolution of diurnal convection over tropical landmasses. / Correlation analysis shows baroclinic conversion (diabatic generation) accounting for 80% (20%) of vacillation-scale dA/dt fluctuations. Baroclinic conversion contributes considerably (69%) to the synoptic variability only during synoptic-scale APE collapses (23 January to 4 February), illustrating the prominence of resolvable baroclinic processes during these periods.

By defining cyclone growth as a logarithmic change in absolute vorticity of the surface center, we investigate growth rates of a ten year sample of cold-season (1 October through 31 March) extratropical cyclones in the North Pacific Ocean basin. Strong growth is frequent in coastal regions near Japan and in the East China Sea. In contrast to the conventional deepening rate intensification parameter, frequency distributions of maximum 24-h growth are found to be more Gaussian in character. / To investigate possible physical processes acting to enhance growth, composite sea surface temperature (SST) anomalies are computed for groups of cyclones which deepen maximally in the vicinity of the Gulf Stream and Kuroshio Currents. Comparison between groups of similar deepening revels stronger growth is generally associated with warm SST anomalies and with weaker upper-level support for intensification. Composite latent and sensible heat flux fields computed from observations support this finding.

In this work, we explore the possibility of combining information from the VAD technique, with data from a wind profiler, to obtain the kinematic properties of a linear 2-D wind field. The idea is first tested on an artificial test-bed wind field, and the results show that the procedure is limited by the degree of non-linearity of the wind field over the wind profiler. To overcome this problem, some modifications are made to our procedure, which is then applied to real cases of stratiform precipitation. We find that the non-linearities are important enough such that the linear approximation is only valid out to about 40 km from the radar. These non-linearities also affect the retrieval of the vorticity to the extent that we had to establish a maximum limit on the uncertainty of the vorticity. These results bring to question the assumptions that the non-linearities are relatively unimportant in stratiform cases.

NCEP reanalyses have been used to compute the climatological isentropic potential vorticity (IPV) budget at the 315K level for 39 winters and for those winters with a North Atlantic Oscillation (NAO) as well as winters with a Pacific/North American (PNA) pattern. / The climatology shows two main IPV sources, each being upstream of the two main centers of positive PV on the east coasts of North America and Asia. The results for the winters with a NAO (PNA) anomaly show, in particular, that the mean-winter IPV anomalies associated with these patterns also have upstream sources. The importance of the latter is not as clear as that of the continential sources that maintain the climatological centers. / The mean-winter IPV advection that balances the IPV sources/sinks is composed of the advection by the time-mean flow and by the transient eddies (decomposed into high- and low-frequency components), where the former is the dominant component. The latter are found to produce a negative feedback in that they act to reduce the amplitude of the IPV anomaly. For the NAO anomaly, low-frequency transient advection is more important, while high-frequency transient advection is more statistically significant for the PNA anomaly. Both the high and low-frequency advection have comparable contributions in maintaining the climatological distribution of the stationary eddy IPV.

Using three cameras and two high powered flash lamps, the HYdrometeor Detection and Ranging using stereO-Photography (HYDROP) experiment, we made the first determination of the sizes and positions of raindrops (diameters >0.2 mm) in a significant volume (approx. 10 m3). Analyzing six three dimensional reconstructions---with 11000--22000 raindrops each and with nominal rain rates in the range of 8--15 mm/hr.---we performed the first direct investigation of the continuum limit of a large number of particles in cloud/rain physics. Since the mean interparticle distance was 7--8 cm and the largest sphere has a radius of the order 2 m, our range of scales is only roughly a factor of 10. Nevertheless, we were able to show clear evidence that a non-trivial multiscaling limit was obtained at least for one rain event out of the two studied; the q th moment of the liquid water density in a radius r was found to scale with a nonlinear (but convex) exponent K(q), characteristic of multifractals. In contrast, when the particle positions were randomized, clear convergence to the classical (homogeneous) continuum was observed with K(q) = 0. By raising the raindrop volumes to a power n = 1/3, in order to improve convergence towards the continuum limit, the universal multifractal parameters of this quantity derived from the field of liquid water content were estimated to be alpha = 1.95 +/- 0.1, C1 = 0.05 +/- 0.01 and H = -0.01 +/- 0.01.

This study is concerned with the effect of slantwise convection in general circulation models (GCMs). The approach is through the development of a slantwise convective parameterization scheme (SCPS) and its implementation into version 11 of the third generation GCM of the Canadian Climate Centre for modelling and analysis (CCCma gcm11). / We first study the characteristics of conditional symmetric instability (CSI) in an environment which is also unstable for conditional upright instability (CUI). The results indicate features common to both upright and slantwise convection. This so called slantwise buoyant instability (SBI) possesses two relevant time scales and its horizontal scale can ranges from tens of km up to over one thousand km. / We then analyze the 15-year ECMWF re-analysis (ERA) data to compute the global distributions of convective available potential energy (CAPE) and slantwise convective available energy (SCAPE). We show that the potential for CSI and CUI indeed co-exists over most areas around the globe. Based on the results of the theoretical study and the data analysis, a parameterization for slantwise convection was developed and implemented into gcm11. It was found that the scheme impacts significantly the simulated general circulation by the development of a direct meridional secondary circulation. The results of the 5-year simulations show that the scheme reduces SCAPE and SCAPE residual rs over the mid-latitudes, leading to a weakening of the thermal wind and the strength of the upper-level jets. The largest improvement in the simulated climate however lies in the reduced meridional transient eddy transports of heat and zonal momentum. With the inclusion of the scheme, the eddy transports agree much more favorably with the observational analysis.

Global data and numerical models are used to investigate the climatology and variability in the upper-troposphere and lower-stratosphere where large-scale low-frequency vorticity modes dominate over high-frequency inertia-gravity waves. An algorithm based on the Empirical Normal Mode decomposition technique is proposed as a diagnostic tool. This technique combines dynamics (a wave-activity of linear dynamics defines the orthogonality between modes) with statistics (the wave-activity amplitude of each mode provides a measure of its statistical significance) and has the unbiased ability to capture neutral normal modes as well as singular modes. / First, the algorithm is applied to winter data provided by the National Centers for Environmental Prediction (NCEP) reanalyses and results indicate that most wave-activity is carried by large-scale, eastward-propagating modes centered at mid- and high-latitudes. The phase-speeds of some leading modes are in good agreement with theoretical predictions of linear dynamics. Some mid-latitude modes exhibit properties that can be explained by the theory of quasi-modes---for example, the leading wavenumber-5 mode with a dipolar pressure field near the tropopause, a propagation speed of 12 m s-1 and a decay rate of 3 days. / In a model study, we use data from two dynamical-core experiments of the Global Environmental Multiscale model: one with the forcing proposed by Held and Suarez, later modified by Williamson et al. (called HSW experiment); the other with the forcing by Boer and Denis (BD). Modes and spectra are similar to those found in the NCEP data study, although details depend on the forcing. For instance, wave-energy amplitudes are higher with the BD forcing and an approximate energy equipartition is observed in the spectrum of wavenumber-1 modes in the studies of NCEP data and BD experiment, but not in the HSW experiment. / The HSW forcing has a relatively strong relaxation acting on the complete temperature field, whereas the BD forcing only acts on the zonal-mean temperature letting the internal dynamics alone drive the wave-activity cascade through the rest of the spectrum. This difference seems to explain why the BD forcing is more successful in reproducing the observed atmospheric wave-activity.

The effects of longwave radiation on the development of a small cumulus cloud were investigated by a combination of three-dimensional radiative transfer model as well as slab-symmetric and axially symmetric cloud dynamics models. / A longwave radiation model was developed. The model was first applied to the case of an isolated cumuliform cloud in the form of a cylinder, the results of which illustrate the importance of cloud side cooling. We proposed that the radiatively forced sinking of air at the sides of cumuliform clouds would lead to lower-level convergence and thereby enhances the upward motion in the interior of the cloud. / A study of cloud top cooling rate in stratiform clouds with positive (lifting) or negative (sinking), axially symmetric perturbations shows that a lifting of the cloud top does not increase cooling rate, whereas the peak cooling rate decreases rapidly in depressions. For small perturbation, $ approx$10 m, this may tend to inhibit the growth of negative perturbations. For larger perturbations, $ approx$100 m, changes to the radiative cooling rates within the positive perturbations and next to the negative perturbations may act to sustain the perturbation and promote its growth. / Using a three-dimensional longwave radiation model and a slab-symmetric cloud dynamics model, we demonstrated that longwave radiative cooling substantially enhances the maximum cloud water content. The maximum increase reaches 96%. The total cloud water was also increased somewhat (maximum 20%). / In the initial stage of the development, the augmentation of cloud water content near the cloud top and sides is traced mainly to the direct effect of longwave radiative cooling on cloud microphysics (i.e. radiative cooling reduces the local temperature and hence the saturation water vapor pressure, which leads to additional condensation). / In the mature stage of cloud, the increase of total cloud water content comes from a combination of the effects of radiation on microphysics and dynamics. The cooling from radiation and evaporation produces additional downward motion at the sides of the cloud. The enhanced low-level convergence invigorates the updraft promoting further cloud development. / In the decaying stage, the negative buoyancy produced by cloud top radiative cooling and a higher liquid water load speeds up the decay process in the LW run. / In a sheared environment, shear suppresses convection. In conjunction with horizontal momentum transport, radiative cooling also results in a more negative temperature perturbation and a stronger downdraft on the downshear flank relative to the upshear side. / In an axially symmetric simulation, longwave cooling produces a weaker updraft in the cloud core. This phenomena was attributed to the shorter lifetime of the axially symmetric cloud, and the fact that the cloud side sinking motion can spread around the cloud core instead of being confined to a vertical plane. As a result, a weaker secondary circulation develops which is offset by the negative effects of the cloud top radiative cooling and the higher liquid water load.

The benefits of very high resolution data from scanning and vertically pointing radars in a research and operational environment are investigated. They are used to perform research on four test topics: the study of the melting layer of precipitation, the derivation of vertical wind profiles using conventional radars, high resolution radar hydrology, and a case study of an unusual weather event. The resulting findings include a quantitative description of bright band statistics, the development of a unique wind measuring technique, the quantification of the sampling problem at high resolution, and a description of the complex interaction between stratiform precipitation and small convective cells. These results show the value of these data for research and hence operations. However, one must avoid the temptation to extract more information from these data sets than can be unambiguously retrieved.