A High-Performance Game Engine-GIS-CFD System for Interactive Urban Wind Decision Support

Abstract. Wind simulation plays a critical role in urban design and planning. Most existing workflows rely on professional and general-purpose computational fluid dynamics (CFD) software. Although powerful and versatile, such tools are not designed specifically for the urban planning process, where interactivity and real-time feedback are highly emphasized. To address this gap, we propose a responsive and specialized decision support system tailored for urban wind simulation. The system enhances the decision-making process by offering an efficient workflow that seamlessly connects data preparation, simulation, and visualization steps. Its core contribution lies in the tight integration of a high-performance CFD model, a game engine, and multi-source geographical information system (GIS) data within a unified environment. The system is built in the Unity game engine to ensure efficiency, user interactivity, scene modifiability and intuitive visualization. The spatial extent and scalability of the system is expanded by incorporating data from multiple GIS sources to represent real-world urban contexts. The wind simulation is accelerated using Unity compute shaders to achieve fast, fully parallelized computation while maintaining physical fidelity. The system also includes NetCDF-based file export and a web viewer. In our testing platform, the system achieves 8.15× acceleration compared to OpenFOAM. The simulation output is also validated through a case study. By combining these elements, we construct a Game Engine- GIS-CFD triplet that serves as a robust decision support system. Its efficiency and usability make it well suited for wind-related urban planning.