Flow and dispersion over a cluster of ‘real-world’ tall buildings

We focus on a cluster of tall buildings in the City of London, situated within a compact area approximately 700m in diameter. We uniquely bring together three methodologies to address the impact of densely packed tall buildings onflow and dispersion. We use real-world wind sonic anemometer observations (191 m above ground level) to identify a day with nearly neutral, stationary urban boundary layer flow and determine the integral time and length scales. We usethese conditions to inform wind tunnel experiments (1:500 scale) and full-scale high-fidelity large-eddy simulations (LES) with the corresponding integral time and length scales approximated and synthetic inflow turbulence generationapplied in the LES.We examine the impact on the building cluster wake region and dispersion of a ground-level release by considering multiple wind directions, and use the wind tunnel-scale observations to evaluate the LES. There isgood agreement, with Reynolds number independence of the flow characteristics confirmed under the investigated conditions. The 17 tall buildings exhibit a cluster effect (i.e. acting with unitary effect) on the flow and a ground-leveltracer release plume as they pass through the buildings. Despite the cluster porosity having large variations with wind direction, the cluster-area-averaged turbulent stresses are much less sensitive to wind direction than the dispersivestresses. For the least porous direction (SW), the Strouhal number based on the identified primary vortex shedding frequency, freestream velocity and effective cluster width corrected for porosity, is found to be close to that of anisolated tall building. In the far wake, the wake width increases following an approximate power-law trend with an exponent of 0.5, while the peak velocity deficit decreases according to a power-law with an exponent of about −1.0.The peak velocity deficit remains above 10% of the freestream velocity even at approximately 3 km downstream of the building cluster. The tracer plume width increases more rapidly than a linear trend in both horizontal and vertical directions just upwind of the cluster, but downstream its growth follows a power-law trend with an exponent of approximately 0.5. This novel combination of methodologies in the challenging environment of tall buildings offersimportant insights for a wide range of applications across multiple scales. These findings are becoming increasingly important as urban populations grow and densification trends drive cities to expand vertically.