Transport and dispersion simulation in downtown Oklahoma City and New York City

Abstract

In this video, we showcase two atmospheric dispersion simulations in an Oklahoma City dataset and a New York City (NYC) dataset. These simulations are created using a robust and efficient framework that generates seamless 3D architectural models from overlapping 2D footprints. These footprints with elevation and height information are commonly used to depict various components of buildings in GIS software such as ESRI ArcGIS and urban model synthesis methods, and usually contain small, sharp, and various (nearly) degenerate artifacts due to machine and human errors. In the first part of the video showing a simulation in Oklahoma City, the location is south of the public library in an area where there is a building currently. Two iso-surfaces of 10-4 and 10-5 ppm are shown in green and the brown clouds. The inflow is a westerly wind with a wind speed of 5 m/s at 10 meters above ground level. In the second part of the video showing a simulation in NYC, the location is the Financial District, Manhattan. The simulation assumed a boundary condition for the inflow of a logarithmic profile of 2 m/s with a velocity at 10 meters from the ground. An iso-surface of 10-5 ppm is shown. The final volume mesh produce contains 333 million tetrahedra, and 59 million points. The total time of the NYC simulation, including the initialization time and dispersion, took approximately two days on a high performance computing system running 2048 cores in a CRAY XK6 nodes. In both simulations, the release is continuous.