Large-eddy simulation of stably stratified atmospheric boundary layer turbulence: A scale-dependent dynamic modeling approach

Over land, stable conditions (turbulence dampened by negative buoyancy) are usually characteristic of nocturnal atmospheric boundary layers. Similar stratified conditions are also found in polar regions and often persist for several months during winter. However, in comparison with convective and neutral atmospheric boundary layer turbulence, stable boundary layer (SBL) turbulence has not received much attention despite its scientifically intriguing nature and practical significance (e.g., numerical weather prediction---NWP, and pollutant transport). The overall goal of this research is to seek a better understanding of the dynamics and subgrid-scale physics of stably stratified atmospheric boundary layer turbulence using a combination of theoretical and modeling approaches.; Specifically, a new-generation subgrid-scale (SGS) modeling approach, named as the 'locally averaged scale-dependent dynamic' model is developed and implemented to perform reliable large-eddy simulation (LES) of SBL. This tuning-free SGS model performs remarkably well in simulating moderately stable boundary layers even at coarser resolutions, at which many traditional SGS models spuriously laminarize. In many other aspects (especially in terms of simulating surface-layer properties), our SGS model is found to be superior to traditional SGS models. An LES database is generated and used in conjunction with extensive field observations and wind-tunnel measurements to further explore the inherent characteristics of stable boundary layer turbulence.; In a parallel line of research, we propose a simple yet efficient synthetic turbulence and passive scalar generation scheme based on the concept of fractal interpolation. It is shown that this scheme not only captures the desirable small-scale properties of turbulence, but also has the potential to be used as a SGS closure model in large-eddy simulation.; Various issues explored in this thesis are not only critical to the progress of micrometeorology and turbulence modeling research, but also for the improvement of boundary-layer parameterizations in numerical weather prediction and climate models.