Estimating multi-scale ventilation corridors in complex 3D urban space : a graph-based least-cost path model

Abstract

Cities are becoming denser and taller, posing increasing challenges for optimizing urban ventilation. Accurately identifying Ventilation Corridors (VCs) is therefore critical to alleviate the urban heat island effect, improve urban air quality, and enhance urban planning and design. However, traditional two-dimensional (2D) methods cannot accurately represent complex vertical airflow dynamics, resulting in the impacts of the depth dimension being ignored, especially in compact-high urban environments. To address this issue, we proposed a graph-based 3D Least-Cost Path (LCP) algorithm based on a voxel-based 3D urban model to identify both vertical and horizontal VCs in the complex urban area of Kowloon Peninsula in Hong Kong. Specifically, the optimal major VCs were determined using a shortest-path searching algorithm and validated by Computational Fluid Dynamics (CFD) simulations. Results show that the proposed model effectively identifies high-performance ventilation pathways, where wind speeds within the VCs are about 1.43 times the global wind speed at all resolutions and wind directions. Notably, the number of VCs per unit time decreases exponentially as the resolution becomes finer, with computational throughput dropping from 12,379 paths/s at 100 m resolution to 6246 paths/s at 50 m and 1351 paths/s at 30 m resolution. It indicates the robust scalability of our 3D LCP model to flexibly balance computational efficiency and spatial precision. Overall, the proposed 3D LCP model is promising to more accurately describe urban air flow, thereby supporting microclimate improvement and sustainable urban design in high-density cities.