Development of a computational fluid dynamics model with tree drag parameterizations: Application to pedestrian wind comfort in an urban area

Abstract

We implemented a tree-drag parameterization scheme in a computational fluid dynamics (CFD) model and validated the results against wind-tunnel measurement and large-eddy simulation data using several statistical measures. The CFD model underestimated (overestimated) the pollutant concentrations on the leeward (windward) walls inside the street canyon in the presence of trees, because the CFD model could not resolve the latticed cage or simulate the concentration changes caused by the latticed cage. However, the simulated pattern and magnitude of pollutant dispersion were similar to those in the wind-tunnel measurements. The CFD model generally satisfied the statistical validation indices (e.g., normalized root mean square error, geometric mean variance, correlation coefficient, and proportion of predictions within a factor of 2 of the observations) but failed to satisfy the fractional and geometric mean biases due to underestimation (overestimation) on the leeward (windward) wall. We evaluated the trees' effect on pedestrian wind comfort in an urban area, using wind comfort criteria based on the Beaufort wind-force scale (BWS) to investigate the sensory level with respect to human activities. In the tree-free scenario, BWS values of 4 and 5 (i.e., unpleasant conditions for sitting for long and short periods, respectively) appeared in narrow spaces between buildings, the upwind side of buildings, and unobstructed areas. In the tree scenario, the BWS values decreased by 1–3 grades on the Pukyong National University campus located in the target area, indicating that trees planted on the campus effectively improved pedestrian wind comfort.