大涡模式垂直分辨率在夹卷及示踪物垂直传输模拟结果分析中的作用

基于敦煌野外观测资料和大涡模式,研究了垂直方向不同尺度湍涡对夹卷及示踪物垂直传输的影响,明确了模式垂直分辨率在模拟结果分析中的作用。结果表明:垂直方向上小尺度湍涡对夹卷作用贡献更大,小尺度湍涡较多时夹卷层相对更暖,而夹卷层厚度、夹卷强度和风速变化受垂直方向湍涡尺度影响较小。当垂直分辨率为50 m时,越往夹卷层上部,上升气流和下沉气流分布较多且强度较大;分辨率为10、20和30 m时,夹卷层各高度垂直速度、位温和示踪物浓度分布较接近。另外,垂直方向湍涡尺度对示踪物垂直传输高度影响不大,而对示踪物的空间分布有一定作用。当大尺度湍涡较多且强度较强时,越有利于将高浓度的示踪物向上传输。综合考虑到模式采用较高分辨率模拟时产生的噪音及计算时间等问题,认为模式采用30 m的垂直分辨率,既能较好地模拟出夹卷层平均结构特征,又能模拟出夹卷层湍流的精细分布,是较为理想的选择。

Effect of Large Eddy Vertical Resolution on Analysis of Entrainment and Tracer Vertical Transport Simulation

The effects of different scales of turbulent vortices in vertical direction on entrainment process and vertical transmission of tracers were studied, and the role of model vertical resolution in the analysis of simulation results was clarified by using field dense observation data in Dunhuang and large eddy model (LEM). The results showed that, in the vertical direction, the small-scale turbulence contributes more to the entrainment, and when small-scale turbulences are more, the clamping layer is even warmer. However, the thickness, strength and wind speed of the clamping layer are less affected by the scale of vertical direction turbulent vortices. Over the upper part of the entrainment layer, there are more updrafts and downdrafts with stronger intensities in the 50 m vertical resolution test. The distributions of vertical velocities, potential temperature and tracer concentration at different heights of the clamping layer are similar in the 10 m, 20 m and 30 m vertical resolution tests. In addition, the scales of vertical turbulent vortices affect the tracer vertical transport height slighty, but has a certain effect on the spatial distribution of tracer. When the large-scale turbulences are more and stronger, they are more favorable for the vertical transmissions of high-concentration tracers. Considering the noise generated in the simulation process and the longer computing time during the high resolution simulations, we use the 30 m grid spacing as LEM vertical resolution in simulating, which is an ideal choice. In such a case, the model can simulate not only the average structure of the entrainment layer, but also the very fine turbulence distribution in the clamping layer.