Abstract
Turbulent flow in a densely built-up area of Seoul, South Korea, for 0900–1500 LST 31 May 2008 is simulated using the parallelized large-eddy simulation model (PALM) coupled to a mesoscale model (Weather Research and Forecasting Model). Time-varying inflow that is composed of mesoscale wind and turbulent signals induces different mean flows and turbulence structures depending on time. Sweeps induced by upper flow are distinct for 0900–0910 LST, and strong ejections and weaker sweeps are dominant for 1450–1500 LST at height z = 200 m. To investigate pedestrian wind environment and ventilation, mean wind velocity and turbulent kinetic energy at 2.5 m above streets are analyzed. The reference mean wind speed at z = 600 m continuously increases after 1010 LST. The pedestrian mean streamwise velocity tends to decrease after 1100 LST, although the pedestrian mean wind speed tends to slowly increase. Whereas the temporal velocity variations related to mesoscale wind are distinct in a street canyon and an intersection, the variations induced by mesoscale wind disappear in a dense building area, indicating strong decoupling from mesoscale wind. The velocity ratio of the pedestrian mean wind speed to the reference mean wind speed, representing a measure of ventilation in urban areas, is high on broad streets and at intersections and is low in dense building areas. Vortices in street canyons and winding flows around tall buildings seem to induce high velocity ratio there. The velocity ratio is shown to be linearly proportional to the pedestrian mean streamwise velocity.
