GPU-based simulation of cold air flow for environmental planning
Stephan Nowatschin, Martin Bertram, Christoph Garth · 2006
Simulating the effects of different land-use types regarding flow resistance and cold air production is important for controlling air quality around urban areas. In this paper we present a mathematical model and a simulation method for this problem. This model describes the cold air flow to be composed of two variables. The first is the velocity field which depends on flow resistance and the flow gradient. The second variable is a height field of the cold air which depends on cold air production and advection. To accelerate the simulation and its visualization, it is adapted to run on a GPU(Graphical Processing Unit). Implementing the simulation on fragment shaders makes it possible to combine the height field of the landscape with a color-coded volume rendering of the associated cold-air height. In two passes we compute the cold air height for each time step and render the result to a texture. In a third pass, we render the height field of the landscape using this texure as multi-layered opacity map.