Fast fluid dynamics simulations of the drag effect of trees on airflow distributions

Trees are widely recognized for their effectiveness in regulating urban microclimates through shading, absorption and reflection of solar radiation, and transpiration. However, their drag effect on airflow may influence this regulatory capacity. Incorporating tree source terms related to leaf area density (LAD) and drag coefficient (C_d) into governing equations provides a balance between computational accuracy and efficiency when studying the drag effect of trees on airflow. Nevertheless, conventional simulation methods typically require significant computational time, limiting their practicality. In this study, tree source terms are integrated into the Fast Fluid Dynamics (FFD) method, and the computational performance of three FFD methods (i.e., SLFFD, NIPC, and NSPF) is evaluated for quickly predicting the drag effect of trees on airflow. Results indicate negligible differences between the predictions of FFD methods and conventional numerical simulation methods such as the Pressure-Implicit with Splitting of Operators (PISO) method. At the single tree canopy scale, the computational speeds of NIPC and NSPF methods are about 1.77 and 1.96 times faster than the PISO method, respectively, while the SLFFD method is about 1.50 times faster. When using the maximum time step size and a first-order discretization scheme, the computational speed of the SLFFD method increases to 4.13 times that of the PISO method. In larger computational domains, the improvement in computational speed provided by the FFD methods becomes even more pronounced. In conclusion, the FFD methods coupled with tree source terms significantly improve computational efficiency for predicting the drag effect of trees on airflow.