Monday, 8 January 2018: 9:30 AM
Salon G (Hilton) (Austin, Texas)
Recent advances in atmospheric modeling have demonstrated that it is now possible to resolve the detailed interactions between the atmosphere and the outdoor urban environment (Lundquist et al. 2012 and Tomas et al. 2017) that are necessary for producing short time averaged, “single-realization”, dispersion solutions. These single-realization dispersion solutions have been shown to be critical for atmospheric dispersion applications associated with air sampling network designs, pollutant measurement systems performance, and characterizing the impact of hazardous airborne materials on human health (Bieringer et al. 2014). Here we describe a Large Eddy Simulation (LES) atmospheric and coupled outdoor urban dispersion model implemented on a Graphics Processing Unit (GPU) computer that has been linked to a building interior air exchange model. This system, referred to as the Joint Outdoor-indoor Urban LES (JOULES) system, is a physics-based quantitative modeling system that is being developed to provide high-fidelity simulations of urban and interior pollutant concentrations for use in the testing and evaluation of operational urban emergency response modeling tools and subsequent enhancements to these systems. JOULES provides physically realistic, time varying atmospheric conditions that interact with the urban landscape and influence the dispersion patterns at locations where air exchanges between the indoor and outdoor spaces occur. A key element of JOULES is the computationally efficient GPU-based LES that enables the development of solution ensembles that can then be used to resolve the distributions in material transport and scenario outcomes associated with the complex interactions between the indoor and outdoor spaces. This paper describes validations of the GPU-LES dispersion model solutions to field observations and provides example simulations that illustrate high temporal/spatial resolution urban contaminant transport and dispersion.
![](images/flag.png)
![](images/blue_ribbon.png)