City-scale urban transport and dispersion simulation using geographic information system footprints

A large volume of urban models describing urban objects in major international cities has been re-constructed and become freely and publicly available via software like Arc-Globe and Google Earth. However, these models are mostly created for visualization and are loosely structured. For example, current GIS software such as ESRI ArcGIS and urban model synthesis methods typically use overlapping 2D footprints with elevation and height information to depict various components of buildings. In this paper, we present a robust and efficient framework that generates seamless 3D architectural models from these footprints that usually contain small, sharp, and various (nearly) degenerate artifacts due to machine and human errors. We demonstrate the benefits of the proposed method by showcase an atmospheric dispersion simulation in a New York City (NYC) dataset. Finally, we discuss several examples of visualizing and analyzing the simulated Computational Fluid Dynamics (CFD) data into the GIS for further geospatial analysis.