We present ensemble-based large-eddy simulations based on a lattice Boltzmann method for a realistic urban area. A plume-dispersion model enables a real-time simulation over several kilometres by applying a local mesh-refinement method. We assess plume-dispersion problems in the complex urban environment of Oklahoma City on 16 July using realistic mesoscale velocity boundary conditions produced by the Weather Research and Forecasting model, as well as building structures and a plant-canopy model introduced into the plume-dispersion model. Ensemble calculations are performed to reduce uncertainties in the macroscale boundary conditions due to turbulence, which cannot be determined by the mesoscale model. The statistics of the plume-dispersion field, as well as mean and maximum concentrations, show that ensemble calculations improve the accuracy of the simulations. Factor-of-2 agreement is found between the ensemble-averaged concentrations based on the simulations over a 4.2 × 4.2 × 2.5 km2 area with 2-m resolution with the plume-dispersion model and the observations.

中文翻译：

---

使用局部网格细化格子 Boltzmann 方法在俄克拉荷马城进行实时示踪剂色散模拟

我们提出了基于格子 Boltzmann 方法的基于集成的大涡模拟，适用于现实的城市地区。通过应用局部网格细化方法，羽流扩散模型能够在数公里范围内进行实时模拟。7 月 16 日，我们使用天气研究和预测模型产生的真实中尺度速度边界条件，以及引入羽流扩散模型的建筑结构和植物冠层模型，评估了俄克拉荷马城复杂城市环境中的羽流扩散问题. 执行集合计算以减少由于湍流而导致的宏观边界条件的不确定性，而湍流无法由中尺度模型确定。羽流扩散场的统计数据，以及平均和最大浓度，表明集成计算提高了模拟的准确性。在 2 米分辨率的 4.2 × 4.2 × 2.5 平方公里区域模拟的集合平均浓度与羽流扩散模型和观测值之间发现了 2 的一致性。